cmg2005 • orlando · 2020-02-20 · cmg2005 • orlando ... seasoned veterans,...

CMG2005December 4th - 9th

Gaylord PalmsResort &

Convention CenterOrlando, Florida

CM

G2

00

5•

Or

la

nd

oLetter From the CMG2005 General Chair,

Cathy Nolan

The Computer Measurement Group invites you join us at CMG 2005, the

31st Annual International Conference of the Association of SystemPerformance Professionals. This year’s conference will be held from

December 4-9 at the luxurious Gaylord Palms Resort and Convention Center in Orlando,Florida.

Program Chair, Lin Merritt and his Program Committee have worked hard to pull together anagenda that offers the systems management community a full week of comprehensive trainingby industry recognized experts and IT professionals. This conference offers attendees achance to interface with presenters in a variety of venues including workshops and “how-to”tutorials, as well as user-experience and problem-solving presentation sessions. Thesetechnical sessions continue to be the primary focus of the conference, and we know you won’tbe disappointed by what you see and hear!

While we continue to deliver outstanding technical sessions, the Business ManagementShowcase sessions were so well received last year that we are once again offering them.These sessions offer attendees the opportunity to attend presentations that examine thechallenges of IT measurement and management from a business perspective.

As always, CMG offers a Trade Show accompanying the conference. Visit the exhibition halloften and learn what solutions our exhibitors have available for you today, as well as for 2006and beyond! Along with the Trade Show, on Monday December 5, are the Monday User Groupmeetings where attendees can get in-depth information on vendor product offerings.

And don’t forget about the vendor sessions and the birds of a feather sessions (BOFS) thatare scheduled daily after the regular sessions end. Last but not least, after a full day of trainingand interaction with industry leaders and IT professionals, is the opportunity to meet informallyat PARS (Performance Analyst Relaxation Sessions), where we will all relax and network whileenjoying food and listening to entertainment.

To add just a few words about our meeting facilities -- the Gaylord Palms Resort andConvention Center offers CMG’s 2005 attendees state of the art conference facilities, wherewireless network connectivity will be provided in every session room, and where each hotelroom offers in-room computing systems with high-speed Internet access and a host of hotelservices. We know you’ll find this facility the perfect learning environment!

Please join us for the annual CMG conference. The Conference Committee has worked hardto put this outstanding venue together, and we know that you will reap the rewards of attendingCMG2005 at the conference for years to come!

CMG2005 PRELIMINARY AGENDA

CMG2005 CONFERENCE COMMITTEE AND

BOARD OF DIRECTORSWHAT IS CMG?

The Computer MeasurementGroup, Inc. is the professional

association of techniciansresponsible for the management

of computer systems. It is avolunteer organization whose

primary mission is the educationof its members and the

advancement of the tools andtechniques for computerperformance evaluation.

CMG acknowledges the ownership oftrademarks and registered trademarksthat appear in this publication. They areeach to be regarded as appearing withthe appropriate ® or ™ symbols at firstmention.

PROGRAM HIGHLIGHTS

GENERAL INFORMATION71

4

CONFERENCE AT-A-GLANCE

DAILY SCHEDULE23

The Association of SystemPerformance Professionals

CMG2005 PRELIMINARY AGENDA

SUBJECT AREA DESCRIPTIONS3

2

SESSION CROSS REFERENCE55

SUNDAY WORKSHOPS & SPEAKERS13

MONDAY USER GROUP MEETINGS

& DESCRIPTIONS19

TUESDAY SESSION DESCRIPTIONS30

WEDNESDAY SESSION DESCRIPTIONS40

THURSDAY SESSION DESCRIPTIONS46

FRIDAY SESSION DESCRIPTIONS52

EXHIBITOR DESCRIPTIONS63

KEYNOTE & PLENARY SPEAKERS5

ACKNOWLEDGEMENTS68

CMG2005 SPONSORS

TeamQuest Corporation

Uptime Software

Diversified Software

HyPerformix, Inc.

PerfCap Corporation

Segue Software

HOTEL & REGISTRATION FORMS36

BOARD OF DIRECTORS AND CONFERENCE COMMITTEE

2

PRESIDENT

Jaqui LynchMainline Information Systems

VICE PRESIDENT

Bryan DrakeVITA

TREASURER

Thomas R. Dennison

SECRETARY

Annette KakazuAcxiom

DIRECTORS

Martin D. BrakeIBM Global Services

Claire CatesSAS

Ellen FriedmanSRM Associates, Ltd.

Richard RalstonHumana Inc.

Jerry RosenbergSRM Associates, Ltd.

Barry SokolikCharles Schwab

Dave ThornSunGard Availability Services

CHIEF OF STAFF

Donna FolkertsIBM Global Services

ADVISORY COUNCIL CHAIR

Stephen J. MarksamerAETNA

BULLETIN EDITOR

Thomas R. Dennison

CORPORATE BOOKEEPER

Linda G. Stermer LGS Bookkeeping Services

FOOD & BEVERAGE

Hugh HuntHunt Conference Group

AUDIO VISUAL

Bill HublerAV Dynamics

CMG HEADQUARTERS

OFFICE MANAGER &

SR. CONFERENCE COORDINATOR

Barbara Hazard

PROGRAM COORDINATOR

David Troxel

CONFERENCE COORDINATOR

Michelle Espiritu

IT SPECIALIST / PROGRAMMER

Rob Harrigan

ADMINISTRATIVE ASSISTANT

Kathleen Kinnarney

CMG2005 BOARD OF DIRECTORS

CMG2005 CONFERENCE COMMITTEE

GENERAL CHAIR

Cathy NolanBank of America

ASST. GENERAL CHAIR

Shana BereznayACS

FINANCE

Thomas R. Dennison

MARKETING

Martin D. BrakeIBM Global Services

CONFERENCE ADVISOR

John PilchPerformance Capacity Solutions

ADMINISTRATIVE CO-CHAIR

Shana J. BereznayACS

ADMINISTRATIVE CO-CHAIR

William Jouris

ACCESS CONTROL

Tom MoulderTREX Associates Inc.

AWARDS

Frank BereznayKaiser Permanente

AWARDS ASSISTANT

Alan DeepePerot Systems Corporation

EXHIBIT COORDINATOR

Trevor CoghlanStorageTek Canada

ONSITE NEWSLETTER

Donna FolkertsIBM Global Services

Denise P. KalmBMC Software, Inc.

REGISTRATION COORDINATOR

Jerry RusthovenMBNA Technology Sector

ASST. REGISTRATION

COORDINATOR

Clayton ChingMicromuse

SIGNS COORDINATOR

Stephen DemminPerot Systems Corporation

ASST. SIGNS COORDINATOR

Dan SchwarzUniversity of Wisconsin Hospital

PROGRAM CHAIR

Linwood MerrittBank of America

ASST. PROGRAM CHAIR

Michael Salsburg, PhDUnisys Corporation

BOF COORDINATOR

Ronald KaminskiSafeway Inc.

INVITED SPEAKERS /

SHOWCASE CHAIR

Annie Shum, PhDBMC Software

MENTOR CHAIR

Margaret GreenbergGrunberg Haus

SUNDAY WORKSHOPS

Dr. Bernie DomanskiCity University of NY-CSI

VOLUNTEER COORDINATOR

Joan Arnold-RoksandichHitachi Data Systems

VOLUNTEER COORDINATOR

ONSITE

Kathy Steffens

SUBJECT AREA CHAIRS

Mainframe (z/OS)Ivan L. GelbGelb Information Systems

Unix / LinuxRick LebsackIBM

StorageMark B. FriedmanDemand Technology

Alan SchulmanUnisys

Fundamentals / CoreCompetencyWilliam Jouris

Network / InternetSidney W. SobermanH W Wilson

WindowsMark B. FriedmanDemand Technology

Michael Salsburg, Ph.DUnisys Corporation

Hot TopicsTim R. NortonVisa USA

DE

CE

MB

ER

• 4-9

SUBJECT DESCRIPTIONS

3

These are the areas of expertise that CMG builds, fosters and is chartered to advance:

MAINFRAME (Z/OS)

The mainframe environment today has evolved from MVS and VM to Z/OS. Everpresent and still mission critical to businesses worldwide are: CICS, IMS, DB2, Batchand WLM. Capacity on Demand, Software Licensing and zAAPs are still moreexamples of the mainframe’s complex and ever-changing environment

UNIX/LINUX

The growth of the UNIX and Linux continues at a rapid rate, and on many platformsincluding VM, IBM, Sun, HP, Unisys and Intel. This subject area covers everything, frommanagement and chargeback to I/O performance.

STORAGE

Storage can be defined as one small set of tape drives, one disk subsystem or acomplex, geographically dispersed, networked system. This topic is inclusive ofwherever information is stored, and encompasses NAS, SAN, iSCSi, Fibre Channel,Backup, Recovery, Storage Performance Council and SNIA as well.

FUNDAMENTALS/CORE COMPETENCY

Fundamentals and core competency are topics that are the cornerstone to all that theCMG is about. Capacity planning, modeling, forecasting, simulation, softwareperformance engineering, chargeback, data analysis and performance tuning are alltopics that are key to successfully measuring and managing computer and storageresources.

NETWORK/INTERNET

The power of connectivity has eclipsed the power of the processor, leading to suchincredible phenomena as the Internet. All the components of the Internet, from hubs,routers and switches to the systems and architectures (DNS, IP, proxies, redirection,content distribution and delivery, etc.) that make up the Internet are covered in thissubject area.

WINDOWS

Exchange, .NET, servers and desktop environments are all part of mission criticalapplications in today’s business organizations. The measurement and management ofMicrosoft’s tools and applications is a key component of this subject area.

HOT TOPICS

This subject area introduces and offers a preliminary evaluation of hot topics andemerging technologies including but not limited to virtualization, server consolidation,blade servers, grid computing, compliance and RFID.

CM

G2

00

5•

Or

la

nd

oPROGRAM HIGHLIGHTS

4

CMG2005 Conference Program Highlights

CMG 2005 has built on the successful changes made for recent CMGconferences. The conference will continue to offer three session lengths:“standard” 60-minute sessions, “extended” 90-minute sessions, and “short”sessions (around 30 minutes) within longer session slots. Levels of detailrange from introductory to technical, and sessions are designed to attractseasoned veterans, capacity/performance rookies and managers withcapacity and performance responsibilities or interests. Within each subject

category, you will find tutorials, discussions on tools and techniques, and user experiences withcommon performance and capacity issues.

The session lineup for CMG 2005 is outstanding. The following outline highlights anotherexciting culmination of the event of the “capacity/performance” year.

Keynote and Plenary Sessions – Our keynote speaker, on Tuesday morning, is Dr. ManiChandy of the California Institute of Technology. Our plenary session, on Wednesday morning,is Amy Wohl of Wohl Associates.

Subject Areas and Sessions – Subject Area categories continue to be fine-tuned as CMGbuilds on the success of CMG 2004. Due to continuing cross-platform expertise and interest ofthe capacity planning and performance attendees at CMG, subject areas have beenreorganized as Mainframe (z/OS), Unix/Linux, Storage, Fundamentals/Core Competency,Network/Internet, Windows, and Hot Topics. Important topics within these categories includesuch subjects as load testing, ITIL, data encryption, virtualization, and Software PerformanceEngineering.

Showcase - The Business Management Showcase returns with another excellent lineup of topexecutives to give you updates on technology directions. This track of four consecutivesessions will be offered on Tuesday.

Workshops – CMG 2005 offers another excellent slate of workshops, with seven concurrentmorning and afternoon sessions on Sunday.

Monday User Groups – CMG will continue to offer the traditional user group meetings onMonday, where users of commercial products will hear about user experiences and newproduct updates.

Invited Speakers – CMG 2005 will continue to invite industry leaders in technology and thecapacity/performance fields to give presentations and serve on panels. These offerings insurethat this year’s conference will offer the best “one stop shopping” for capacity and performancecontent, and cover a wide range of capacity planning, performance, architecture, hardware,and software topics.

Don’t miss the conference with content! I hope to see you in Orlando in December 2005!

Lin MerrittCMG 2005 Program Chair

DE

CE

MB

ER

• 4-9

TUESDAY 8:00 AM - 9:00 AM KEYNOTE SPEAKER

K. Mani Chandy

California Institute of TechnologySimon Ramo Professor of Computer Science

Sense and Respond Systems

Sense and respond systems sense, and then respond, to opportunities and threats. A senseand respond platform can be configured to develop specific sense and respond applications.Sense and respond applications arise in homeland security, healthcare, finance, supplychains, energy, environmental protection, security and - most importantly for this talk - themanagement of IT infrastructure. The CMG community can make a profound impact on thespace of sense and respond applications because this community has expertise in therelevant technologies and mathematics: CMG papers deal with measurement ofasynchronous events, statistics, probabilistic models, information fusion, and real time. Senseand respond applications are increasingly important to society, and CMG has a major role toplay. This talk will survey the field of sense and respond applications; identify applicationsspaces that are critical for society; survey the technologies and software architectures usedin these applications; describe fundamental problems in experimentation, systems design andtheory; explore the rate of growth of this space; show how the CMG community's experienceis directly relevant; and discuss experience with developing applications from platforms.

This year's keynote speaker is Dr. K. Mani Chandy, the Simon Ramo Professor of ComputerScience at the California Institute of Technology.

Dr. Chandy got his Ph.D. from the Massachusetts Institute of Technology in ElectricalEngineering, a Masters from the Polytechnic Institute of Brooklyn, and a Bachelors from theIndian Institute of Technology, Madras.

Dr. Chandy has worked for Honeywell and IBM. From 1970 to 1987, he was in the ComputerScience Department of the University of Texas at Austin, where he served as chair. He hasbeen at Caltech since 1987 where he served as executive officer of the Computer ScienceDepartment. He has served as a consultant to a number of companies and is a co-founder ofiSpheres Corp.

Dr. Chandy is a member of the National Academy of Engineering. He received the IEEE KojiKobayashi Award for Computers and Communication and the A.A. Michelson Award from theComputer Measurement Group and he was a Sherman Fairchild Fellow.

Dr. Chandy's research is on distributed computing. His current work is on building an event-directed architecture as a thin layer on top of service-oriented architectures. He is working onsense and respond applications dealing with managing crises. He has published three booksand over a hundred papers on distributed computing, verification of concurrent programs,parallel programming languages and performance models of computing and communicationsystems.

5

PLENARY SPEAKER WEDNESDAY 8:00 AM - 9:00 AM

6

Amy Wohl

Wohl AssociatesPresident of Wohl Associates, Editor and Publisher of Amy D. Wohl's Opinions

The China Era

The China Era: China is on the rise. Sometime in this century America will need to recognize it as (at least) an economicSuper Power.

But what does China’s evolution from an enormous, largely agrarian, economy to an international investor, an educationalpowerhouse, and a manufacturing giant have to do with the business of information processing? The answers are at oncecomplex, disturbing, and exciting. It is one part scary (we’ll loose jobs, even very skilled jobs to Chinese working remotelyfrom the U.S.); one part challenging (they are educating their students much better in science and engineering than we are);and one part empowering (they are eager investors in the U.S. and excited to partner with U.S. firms on projects in China).

Don’t just rerun the Indian Outsourcing tapes. In fact, Western companies are already looking to move some manufacturingwork out of China, as increased labor costs (and the overhead of doing business remotely) makes them too expensive.Instead, be prepared to look at the facts. We’ll bring lots of facts, war stories, and challenging questions.

The Chinese symbol for crisis includes the notion of both danger and opportunity. This session will look at both what aChinese Powerhouse might demand and of the partnering and collaboration opportunities it will enable.

Amy D. Wohl is President of Wohl Associates, a consulting firm located in Narberth, Pennsylvania which consults on newand emerging technology and new market formation. Wohl Associates’ current interests include all types of personal andgroup software, the Internet, and information appliances. She is well known for her past and on-going interests in officesoftware, groupware, and speech processing. The firm provides services on strategic planning, marketing strategy,marketing research, and training to developers of information systems, hardware, and software.

A noted expert on the computer market, its products and dynamics, Mrs. Wohl has served as an expert witness and legalconsultant on many occasions.

Mrs. Wohl is Editor and Publisher of Amy D. Wohl’s Opinions, a weekly electronic newsletter. She previously edited TheTrendsLetter and The Wohl Report. She also maintains a weblog at http://amywohl.weblogger.com., is a contributing editorto Spectrum Middleware, and a frequent contributor to the trade and general business press on the Internet, software,computing, computer trends, and technology. Mrs. Wohl is a frequent speaker at industry conferences and seminars in theU.S., Canada, Europe, South America, Africa, Asia, and Australia. She is especially well known for her keynotes on thefuture.

She has served as a board member and advisor to several corporations. Currently, she is a principal and advisor toInteractive Intelligence, an Internet Design firm, and a member of the Board of Advisors of Bachow Investments. She alsoserves as an advisor to a number of new technology start-ups.

Mrs. Wohl was named one of the top 100 Women in Computing in 1994. She was program chairman of the first OfficeAutomation Conference. She is past president of the Office Systems Research Association and served on the board ofWomen in Information Processing. She was the founding president of the Philadelphia Chapter of IWP. She is a member ofOSRA, ACM, SIGOA, and SIGPC. Mrs. Wohl is the recipient of the Augusta Ada Lovelace Award from the Association ofWomen in Computing.

Mrs. Wohl received a B.A. in Economics from LaSalle College and an M.A. in Economics from Temple University, where shewas an N.D.E.A. Doctoral Fellow. Mrs. Wohl currently teaches in the Masters program in Organizational Dynamics at theUniversity of Pennsylvania. She has been a guest lecturer at many colleges and universities, including the Wharton Schoolof the University of Pennsylvania, Harvard University, Oklahoma State University, Temple University, and the University ofCalifornia at Berkeley.

The Business Management Showcase returns with another excellent lineup of top executives to give you updateson technology directions. This track of four consecutive sessions will be offered on Tuesday.

TBD 9:15 AM - 10:15 AMVice President, Product Management and Product Marketing, ESM Division at Computer Associates Are Integrated Management solutions the recipe for an invisible and secure IT infrastructure?

As businesses today are demanding greater flexibility and efficiency, we see that the IT Organizations aredemanding the management software providers to deliver a complete, integrated and open solution for policy-driven, adaptive Services Oriented Infrastructure (SOI) that can expose assets as composite IT services and enablea contextually relevant view of business processes. In this presentation, we will discuss how we could optimize thedelivery of business services through an invisible and secure infrastructure that can transparently adapt to changesin business conditions. We will also highlight how Management solutions that are integrated across the IT stack arebetter suited to achieve the vision of an invisible and secure infrastructure.

Glenn O'Donnell 10:30 AM - 12:00 PMPrincipal Product Marketing Manager at EMCMaking Sense of the Performance Riddle

Service performance is a “riddle wrapped in a mystery inside an enigma” to the extreme majority of IT organizations.To truly understand this riddle, one must navigate a complex labyrinth of interrelated technology and businesscomponents. New composite applications and distributed services exacerbate this scenario. No human is capableof comprehending all of this complexity, so automation technology, process, and standardization are the keys tocontrolling our environments and to leveraging this complexity for competitive benefit. Some early successes proveit is possible, albeit difficult, as cultural shifts never come easily. Beware of grand promises of automated simplicity.Such nirvana remains many years from reality, if ever.

Herb VanHook 2:00 PM - 3:00 PMVice President at BMC SoftwareOpportunities and Challenges of the SOA World

The loosely-coupled application architectures emerging as the preferred new wave of integration enable flexible andadaptable business models, but come with their own challenges when mapped to the traditional management andoperational process disciplines. The promise of Service-Oriented Architectures presents alignment opportunitiesbetween technology organizations and their business counterparts, as a closer mapping between real-worldbusiness services and application components is realized. As companies move to a "flow computing" paradigm withthe adoption of Web services, many of the lessons learned over the years in running high-performance technologyenvironments will carry forward, but many of them will have to be adapted to the realities of these new architectures.This session will address the imperatives that Service-Oriented Architecture "run-time" brings to the forefront.

Eric Pulier 3:45 PM - 4:45 PMExecutive Chairman at SOA Software, IncLearning to Play Nicely

By their very nature, Service-Oriented Architectures (SOAs) tend to blur traditional boundaries in largeorganizations. In the current environment, most IT initiatives involve the coordination of such distinct groups assoftware development, network operations, security, architects, as well as line of business managers. Not soanymore. With the SOA potentially creating reusable software code that must be accessed dynamically bycomposite applications, both inside and outside the firewall, the traditional roles and responsibilities of IT have beenforever changed, or even erased. In this presentation, technology entrepreneur and SOA visionary Eric Pulier willexplore the ways in which the SOA pushes the boundaries of IT and look at how large organizations can teach theirdisparate IT groups to “play nicely” and evolve into a successful SOA culture.

7

PROGRAM HIGHLIGHTS

BUSINESS MANAGEMENT SHOWCASE TUESDAY

PROGRAM HIGHLIGHTS

8

SUBJECT SESSION AUTHOR TITLE

Core 301 K. Mani Chandy Sense and Respond Systems

Core 316 TBD Are Integrated Management Solutions the Recipe for an Invisible and Secure IT

Infrastructure?

Hot 325 Dr. Toufic Boubez Policy Driven SOA

Core 326 Glenn O’Donnell Making Sense of the Performance Riddle

Hot 336 Herb VanHook Opportunities and Challenges of the SOA World

Hot 346 Eric Pulier Learning to Play Nicely

Win 401 Amy D. Wohl The China Era

zOS 412 Robert R. Rogers How Do You Do What You Do When You’re a CPU?

Core 413 Peter Sevcik The Application Performance Index (Apdex) Standard+

Hot 415 Nick Gall Reflection: The Next Big Thing After SOA

zOS 422 Glenn R. Anderson WebSphere Application Server for z/OS Version 6 - Performance Tuning

Core 431 Dr. Jeffrey P. Buzen Workload Characterization for Parallel Processing Environments

Win 433 Kevin Kline Performance Baselining, Benchmarking and Monitoring for Microsoft SQL Server

Hot 435 Ken Traub Radio-Frequency Identification at Enterprise Scale

zOS 501 Cheryl Watson z990 Performance and Capacity Planning Issues

Net 505 Bernie Davidovics Measuring End User Response Time - With a Passive Network Probe

Net 515 Laura Knapp Introduction to Simple Network Management Protocol (SNMP) Version 3

zOS 521 Kathy Walsh The Mechanics of Developing a High Quality Capacity Plan

Net 535 Chris Loosely Measuring Actual and Perceived User Experience on the Web

Hot 536 Richard Soley The Model-Driven (R)evolution

Hot 543 David R. Morley CMG: The Early Years

INVITED SPEAKERS

SUBJECT SUBJECT AREASAREAS

Core = Fundamentals / Core Competency Hot = Hot TopicsITIL = ITILzOS = Mainframe (z/OS)

Net = Network / Internet Stor = Storage*nix = Unix / Linux Win = Windows

From analysts to industry-recognized experts in the fields of performance and capacity, CMG is the conference that putsattendees in direct contact with highly experienced individuals. CMG sessions provide attendees with the opportunity tonetwork with and learn from the best the industry has to offer. CMG’s Invited Speaker track includes some of the industry’smost recognized visionaries.

PROGRAM HIGHLIGHTS

9


ITIL 313 Chris Molloy ITIL Capacity Management Deep Dive

zOS 322 Kenneth D. Williams MVS Application Performance Management

ITIL 327 Avtar Dhillon Achieve IT Agility by Integrating SOA with ITIL Based BSM

Hot 334 Peter J. Weilnau Measuring Up for Server Virtualization

zOS 342 Don Deese Introduction to zSeries Application Assist Processor (zAAP)

Hot 411 Dr. Jeffrey P. Buzen SOA and the Social Order - City Planning, Post Office & Business Protocols

*nix 421 Jaqui Lynch Linux Performance Tuning

Core 425 Jack B. Woolley Want to Know WHY Response Time is So Long? Listen to the Wire.


Core 437 Irving Smith Wholesale Distributed Capacity Planning

Hot 504 Cathy Nolan Encryption Primer: An Introduction to Data Protection

Core 507 Denise P. Kalm Survivor - The Corporate Jungle

Core 514 Linwood Merritt Closing the Gaps – Understanding Capacity Summarization

Core 517 James A. Yaple Benchmarking 101


*nix 523 Serge Tessier Capacity Planning for UNIX System Metrics

Net 525 Nalini J. Elkins One-Minute TCP Stack Analysis

*nix 527 Robert F. Patterson Capturing System Data Using Native Commands

*nix 532 Rick Lebsack Facilitated Discussion: Future of the Performance Field 2005

*nix 533 James Holtman Visualization of Performance Data


zOS 604 Dr. Bernard Domanski Speaking SOA and Web Services: .NET and the Mainframe

Core 623 Frank Bereznay Making Your Web Portal a Dynamic Website

ROOKIE GUIDE

Suggested Sessions for New Conference Attendees

The following is a set of suggested sessions for attendees new to the national CMG conference, and for those that have notbeen at the the national conference in several years. These sessions were chosen primarily on the basis of their content beingtutorial in nature. In some cases the speaker is well-known within CMG circles, and has consistently delivered high qualitymaterial in a very entertaining way.

CM

G2

00

5•

Or

la

nd

oSATURDAY & SUNDAY CLASS

10

Computer Performance Evaluation

for MVS Mainframe Systems

by Barry Merrill

With the cooperation of the Computer Measurement Group, Dr. Barry Merrillwill offer his normally-three-day-class, “Computer Performance Evaluationfor MVS Mainframe Systems”, tuition-free, taught in two very-long-days,Saturday and Sunday, from 8:30 am to 7:30 pm, December 3 and 4,preceding the CMG2005 Conference in Gaylord Palms conference hotel.

This class is open to technicians who use or plan to use MXG or ITSV/ITRMsoftware. The class description, registration forms, and links to the hotelreservation form and CMG information are in the CLASSES frame atwww.mxg.com. You can leave home on Friday, take advantage of over-Saturday-night and attend both the MXG class and the CMG Conference,all in the same hotel, and at the CMG Conference hotel room rate!

DE

CE

MB

ER

• 4-9

SUNDAY WORKSHOPS

11

CMG2005 SUNDAY WORKSHOP SCHEDULE

BREAKFAST 7:30 AM – 8:30 AM

MORNING WORKSHOPS 8:30 AM – 12:00 PM

LUNCH 12:00 PM – 1:00 PM

AFTERNOON WORKSHOPS 1:00 PM – 4:30 PM

Morning Workshops

Robert D. AndresenMonitoring Linux Hands-on Lab

Adrian Cockcroft & Mario JauvinCapacity Planning and Performance Monitoring with Free Tools

Mark FriedmanScalability and Performance of .NET Framework Applications

Ivan GelbMaximizing Performance on z/OS

Curtis HrischukCapacity Sizing and Performance Engineering of the Mobile Enterprise

Henry S. NewmanI/O Performance Analysis and Tuning: From the Application to the Storage Device

Cathy A. Wright

ITIL - What's IT All About Then?

Afternoon Workshops

Dr. Toufic BoubezBest Practices for Building a Secure and Manageable SOA

Dr. Alexandra FenzlIT Executive Dashboards - Measuring the Essential

James P. HoltmanPerformance Analysis Using R

Randy KernsStorage Strategy Workshop

Dr. Odysseas I. PentakalosTuning the Performance of Java and J2EE Applications

John P. PilchCPE as MIS or Information for Making Decisions

Jack ProbstITIL at Warp Speed

Please note that each workshop is only presented once and will not be repeated.

CM

G2

00

5•

Or

la

nd

o

12

Robert D. Andresen

MQSoftware Inc.

Morning only

Monitoring Linux Hands-on Lab

Linux is gaining interest as a solution acrossmany hardware platforms: x86 based machines,Sun and Apple proprietary hardware and IBMzSeries platforms. But once applications areported to an open source operating system whatoptions are available to monitor theirperformance and availability? This in-depthhands-on lab will provide the unique opportunityfor you to learn while doing and covers nativeLinux solutions for monitoring performance andcollecting statistics for capacity planning. lookingat tools ranging from real-time monitors throughthose that can build a database of historicalsystem performance Students will need to bring aat least a Pentium laptop with at least 128Meg ofmemory that can boot from the CDROM. Abootable Knoppix CD will be provided to do theworkshop exercises. Knoppix will make nochanges to the laptop. Recommended audience:Linux beginners through intermediate.

Robert Andresen is a Systems Engineer withMQSoftware. He has been working inperformance management and monitoring forseveral software companies since 1997. He hasbeen working with Linux since 1995. He is a co-author of the IBM Redbook: Linux on IBM@server zSeries and S/390: SystemManagement. He holds a degree in Mathematicsfrom the Illinois Institute of Technology. AtMQSoftware he supports QPasa and QNami,which provide business transaction assuranceand middleware management. His areas ofexpertise include z/OS, UNIX/Linux, CICS, DB2,the Websphere family, and Networking.

Dr. Toufic Boubez

Layer 7 Technologies

Afternoon only

Best Practices for Building a Secure

and Manageable SOA

Flexibility is one of the major goals of ServiceOriented Architecture (SOA) and the words"loose coupling" have become the newarchitectural mantra. But what does it actuallymean to build a true loosely coupled architecture,and how do you go about doing it? This workshopis intended to provide a relatively comprehensivecoverage of Web Services for architects and ITexecutives. Starting with the original vision andbusiness drivers for SOA and Web services, theworkshop will introduce examples that illustratethe issues that companies face when buildingreal world SOAs. Some of the issues that will beaddressed are: How do you decouple yourbusiness logic from your infrastructure? How doyou implement security, identity, reliability andnon-repudiation in a loosely coupled world? Howdo you establish and enforce policies that allowyou to manage your SOA deployments? Theworkshop will cover some of the mainspecifications such as WS-Security, WS-Trust,WS-Policy and SAML.

Dr. Toufic Boubez is a well-respected Webservices pioneer. Prior to co-founding Layer 7Technologies, he was the Chief Architect for WebServices at IBM's Software Group and drove theirearly XML and Web Services strategies. He co-authored the original UDDI API specification andrepresented IBM on several standards bodies,including the ebXML initiative. Toufic alsomanaged some of IBM Global Services' largeste-business engagements. He is a sought-afterpresenter and has chaired many XML and Webservices conferences. Toufic is also activelyinvolved with various standards organizationssuch as OASIS and WS-I, and is the co-authorof the WS-Trust and WS-SecureConversationspecifications. He is the author of manypublications and three books, including "BuildingWeb Services with Java". In 2002, InfoWorldnamed him to its "Ones to Watch" list. Touficholds a Master of Electrical Engineering degreefrom McGill University and a Ph.D. in BiomedicalEngineering from Rutgers University.

SUNDAY WORKSHOPS

SUNDAY WORKSHOPS

Adrian Cockcroft

eBay Inc.

Mario Jauvin

MFJ Associates

Morning only

Capacity Planning and Performance

Monitoring with Free Tools

Computer system and Network performance data collection,analysis, modeling and capacity planning on any platform (includingbut not limited to Solaris, Linux, Windows, MacOS X) using bundledutilities and freely available tools such as Orca, BigBrother,OpenNMS, Nagios, Ganglia, SE Toolkit, R, MySQL and PDQ.Overview: Capacity planning and performance management toolshave been commercially available for many years. A new generationof freely available tools provide data collectors and analysispackages. As the underlying computer platforms and networkdevices have evolved, they have added improved data sources andhave bundled free data collectors. Several open source andfreeware projects have sprung up to collect and display cross-platform data, and with the advent of highly functional free statisticsand modeling packages (and free databases) comprehensiveanalysis, modeling and archival storage can now be assembledwithout paying per-node license costs and with less integrationeffort and lower support costs than many commercial tools. Freetools are of special interest to sites with very diverse mixes ofsystems, very large sites where licensing costs become prohibitive,and sites replacing a few large single systems with many more lowcost horizontally scaled systems.

- Overview of capacity planning processes

- Data collectionSE toolkit and orcallator for SolarisExtended system accounting for Solarisprocollator for LinuxSNMP and other network data

Windows data collection

- Logging, availability and alert generationopenNMS, NagiosBig Brother, Big Sister

- Data storageorca for data storage with RRD

(R and MySQL data storage schemas - covered in R tutorial)

- Data display(R based plotting and graphing - covered in R tutorial)

gnuplot, Orca, OpenDX (3D data visualization), Openoffice Calc, Excel

- Data analysis and modeling(R based statistical analysis - covered in R tutorial)

PDQ based queuing model analysis

- LicensesTypes of license and restrictions on useSpecific licenses used by products mentioned

Dr. Alexandra Fenzl

easyplex Software GmbH

Afternoon only

IT Executive Dashboards - Measuring the

Essential

Under budgetary and competitive pressure, CIOs and CTOs areincreasingly seeking better ways to tack IT projects and resources,justify their development cost and improve quality of output.Contrarily to the past, IT departments are required to align with thebusiness side of the organization and contribute to highest levels ofperformance. And suddenly, engineers and technology specialistsare on the hot seat, feeling unsure about their destiny andcompetence - all the metrics and various dashboards availablethrough traditional software project and development tools, don´tprovide the proper information anymore. Top management is askingfor hard facts – performance and quality indicators - that theythemselves don´t know how to define. All they know is that simplefunction point analysis or a count of LOC won´t do the trick.

So there you have it - even an infinite number of metrics that can beapplied to various parts of the software development cycle, don´thelp much with the measurement CIOs and IT Project managersare charged with.

Together with clients, easyplex Software GmbH has launched aninitiative to face those challenges and establish Software Qualityand Development Performance Criteria. Based on a comprehensivemetrics catalogue, we developed a central dashboard to presentreal-time status of those indicators and underlying metrics totechnical as well as non-technical top executives.

In an effort to not put another burden on CIOs for having to acquireyet another expensive tool and possibly having to replace some oldtools (which we know is always another hurdle to overcome), thedashboard is designed according to open standards and easilyintegrates/captures the necessary data from underlying tools, suchas PVCS Trackers, Bugtrackers, etc. The solution is also meant toeliminate old and tedious Excel sheets, error-prone as we know,and most of the manual work that had to be performed when amanagement report was due. The dashboard is published to awebsite and is available 24 hours for any management executivewho wants to take a look at the technology team´s performance,customer satisfaction and/or product stabilization/quality metrics.Red, yellow and green lights quickly indicate the dangerous andmost critical areas. For the more technically experienced, drill-downgraphics reflect the underlying data and more detailed facts aboutan issue/risk.

The tutorial will introduce CIOs/Project Managers and TopExecutives to the dashboard, will offer a case study of one ofeasyplex customers using the solution, and provide hands-onexperience with Software development and quality metrics, discussthe use of a Software Project Balanced Scorecard and potentialmissing parts/tools within or around the dashboard.

13

SUNDAY WORKSHOPS

14

Mark Friedman

Demand Technology

Morning only

Scalability and Performance of .NET

Framework Applications

The Microsoft .NET Framework is a comprehensive set ofapplication development and deployment technologies that istightly integrated with a broad range of Microsoft serversoftware products, including the IIS Web Server and MS SQLServer, This ½ day workshop focuses on the scalability andperformance of .NET Framework applications, from design,development, testing through to deployment. Using thediscipline of Performance Engineering as a conceptualframework, the workshop features a variety of practicaltechniques to assist application architects, seniordevelopers, and system and database administrators withresponsibility for designing, building, and deploying .NETFramework applications that will meet or exceed yourorganization’s performance requirements. The workshop willfocus on the architecture of.NET Framework applications,concentrating on ASP.NET web forms and ADO.NET data-driven design and development, and the use andinterpretation of the measurement data that is available for.NET applications.

Mark Friedman is the founder and President of Demand TechnologySoftware, headquartered in Naples, FL. The company developstools for Windows performance monitoring and capacity planning.He is the author of the Windows Server 2003 Performance Guide,a volume in the Windows Server 2003 Resource Kit, published byMicrosoft Press in 2005. His earlier book, Windows 2000Performance Guide, was published by O’Reilly Associates inFebruary, 2002.

Mr. Friedman's experience in commercial Information Technologyspans twenty-five years with Fortune 100 corporate data centers,government, hardware vendors and commercial software houses.His previous experience includes senior technical and managementroles at Datacore Software, Landmark Systems, Morino Associates,and StorageTek. He is a recognized expert in computerperformance and storage management. He was a regularcontributor to Enterprise Systems Journal and is in demand tospeak at Computer Measurement Group, SHARE, GUIDE, StorageManagement User Groups, the RAID Advisory Board, and meetingsof other professional organizations. Mr. Friedman's trainingseminars, lectures and published work are highly regarded for theirtechnical quality and depth, and he is esteemed for his ability tocommunicate complex technical topics in plain, concise terms. Heholds a Master's degree in Computer Science.

Ivan Gelb

Gelb Information Systems Corp.

Morning only

Maximizing Performance on z/OS

Many of you are now being evaluated and judged based onhow effectively you meet service level requirements anddeliver your services within your organization. In this one dayseminar you will learn proven best practices about how to setup, customize, report, and analyze the performance andcapacity of z/OS and its major subsystems. In oneinformation filed day, you will develop your skills to: Tune thez/OS environment beginning with the Workload Manager(WLM) goal mode service definitions and ending with specificexamples of CICS Transaction Server and DB2 performancemanagement. Effectively collect and analyze data andreports from Resource Measurement Facility (RMF) andSystem Management Facility (SMF). Quickly identifyperformance bottlenecks and reduce or eliminate theireffects. Create effective procedures for monitoring andreporting system performance and capacity utilization.Adjust the system options to yield optimum performance at aminimum total cost for hardware and software. All the latestzSeries hardware and software configuration alternativesand major issues affecting performance will be covered, andyou will develop your understanding about how a z/OScomplex can be a truly on-demand service provider. Theareas covered will also include: Intelligent Resource Director(IRD), variable Workload License Charges (vWLC), thezSeries Application Assist Processor (zAAP), and others.Registered attendees are encouraged to either bring to theseminar or email to us compressed files of system definitionparameters and any performance or capacity managementreports of issues they are currently being challenged by.Please include a description of the issues you are trying toresolve. We will not leave your questions within our subjectareas unanswered.

DE

CE

MB

ER

• 4-9

SUNDAY WORKSHOPS

15

James P. Holtman

Convergys

Afternoon only

Performance Analysis Using R

Performance analysis of a computer systemrequires the collection of a number of metricsrelating to the operating system and businessevents that the application is processing. Thisinformation can take many forms:database records, flat files, XML, etc. Once thisdata is collected it needs to be analyzed andcorrelated in order to understand what thesystem is doing.

There are a number of products in themarketplace which can do this (EXCEL, SAS,SPSS, Java, C++, etc.), but this workshop willfocus on the use of R (a statistical programmingenvironment) to process this information. R is'freeware' and provides the ability to handle largeamounts of data (matrices with millions of rows)and can perform most calculations in seconds onthis data. It has a powerful graphics environmentthat allows the visualization of the information invarious ways (PDF, postscript, JPEG, PNG, etc.).It also provides interfaces to relationaldatabases, files and XML encoded data. R isprimarily an environment that supports statisticalanalysis, and provides a very comprehensive setof routines to do linear/non-linear modeling andregression analysis. R is a programminglanguage/environment that supports objectoriented development.

The workshop will provide an overview of R withexamples using data collected from UNIXsystems. The data collection and analysis scriptswill be provided to the students. Attendees candownload the R software from http://www.r-project.org/ so that they can also follow along inclass.

James has worked in the computer field for over 42years and spent the last 30 on performance relatedissues. He is a Bell Labs Fellow for his work inarchitecture validation and for creating large scale OSsystems for the Bell System. His work at Convergys isas an internal consultant on performance andscalability, and James teaches an internal course onperformance for their staff.

Curtis Hrischuk

IBM

Morning only

Capacity Sizing and Performance

Engineering of the Mobile Enterprise

The mobile enterprise is here and it introducesnew technology that extends the edge of theenterprise network. It also adds newperformance engineering issues and concerns.The performance characteristics of a mobiledevice (i.e, smart cell phones and PDA’s) differfrom similar devices (i.e., tablet PC) in radicalways because of CPU, memory, and batteryconstraints. The software infrastructure (e.g.,Java Virtual Machine) and applications usetechnologies whose performance implicationsare only beginning to be examined. As well asoperating in an on-line, connected mode, mobileapplication’s operate in a disconnected modethat use an underlying, bi-directionalsynchronization to maintain consistency with themobile data store and the enterprise data store.The communication mediums vary from minuteto minute, ranging from: low to high speed,reliable to sporadic, etc. Capacity sizing forthese class of systems is challenging becausethe workloads are different than conventionaltransactional systems. This workshop presentslesson learned in the performance engineeringand capacity sizing of mobile enterprisemiddleware and applications. Backgroundinformation about mobile technology is alsopresented.

Curtis Hrischuk is a Lead Performance Engineerfor IBM in the area of Pervasive Computingmiddleware products. He completed the M.Eng.(1995) and Ph.D. (1998) degrees from CarletonUniversity, Ottawa, Canada in systems andcomputer engineering.

SUNDAY WORKSHOPS

Randy Kerns

Independent Storage Analyst

Afternoon only

Storage Strategy Workshop

Storage has become a critical asset for corporations and thepurchase, management, and overall strategy is a key decision thatmay have effects on the overall success of the company. Findingthe time to develop a storage strategy is difficult in and of itself butknowing what comprises a storage strategy and what are theimportant elements is something that most storage professionalshave difficulty in developing on their own. There are many issues tounderstand and alternatives to consider that crucial for developingan effective strategy. Without that information there may be a lackof support for a strategic program. A decision that is not in the bestinterests of the company may be made or influenced based uponhigh-level sales techniques without the proper education as to theproblems and solutions for handling information.

This session will present the information that is needed to developa storage strategy for an enterprise. The information is applicableto executives or to IT staffs that have the responsibility to plan tomeet the business demands. A Storage Strategy workbook will beprovided.

The topics to be covered include:

• Understanding the issues and the need for a storage strategy.• Business requirements to consider and risk management. • The components that are part of a storage strategy.• Where and how to acquire information.• Making critical decisions.• Recommendation for proceeding in designing and implementing

a storage strategy.

Randy Kerns is an independent storage analyst and covers storageand storage management software including Storage Area Networkand Network Attached Storage analysis.

He has over thirty years in the computer industry involved in thedevelopment of storage products for both mainframe and opensystems. His background is in product design and development.Randy's education includes a bachelor's degree in computerscience from the University of Missouri at Rolla and a master'sdegree in engineering computer science from the University ofColorado. He has worked for IBM, Fujitsu, as Vice President ofEngineering at the Array Technology subsidiary of TandemComputers and as Director of Engineering for Enterprise Disk atStorage Technology Corporation. Product development that Randyhas been involved in includes both disk and tape subsystems forthose companies.

Randy has made numerous presentations at conferences and is theauthor of many industry articles and white papers.

Henry S. Newman

Instrumental Inc

Morning only

I/O Performance Analysis and Tuning: From

the Application to the Storage Device

This tutorial explores the issues surrounding I/O performanceand tuning. Far too often organizations take an approachthat does not yield the best results when it comes to tuningI/O systems. For example not having the database architect,working with the server file system group and not havingeither of them working with the storage group will ensure thatthe I/O performance will not be optimal, and that likely will notscale. Without I/O scaling given that storage followsAmdahl’s Law, adding more devices will not necessarilyimprove performance.

This tutorial will address the end-to-end performance andtuning issues, and will hopefully allow all groups to worktogether with a common knowledge base.

Mr. Newman has worked in the IT industry for over 24 years.In positions with Instrumental and Cray Research, Mr.Newman has provided expertise in systems architecture andperformance analysis to customers in government, scientificresearch, and industry around the world.

1992-Present CTO Instrumental Inc1981-1992 Cray Research (various positions)

Recent Activities• Ground architect for major defense contractor

for high speed data collection• Architecture team for the Library of Congress

National Audio Visual Conservation Center (8 PByte

per year archive)

• WAN/Shared file system Principle Investigator for large defense contractor

• Disaster Recovery Principle Investigator for DoD HighPerformance Computing Modernization Office multi-petabyte archive

16

SUNDAY WORKSHOPS

Dr. Odysseas I. Pentakalos

SYSNET International, Inc.

Afternoon only

Tuning the Performance of Java and J2EE

Applications

This workshop will provide a comprehensive overview on thedevelopment of high-performance Java-based applications and onthe tuning of existing applications. Particular emphasis will beplaced on enterprise and server-side applications.

The workshop will begin by reviewing the execution process of aJava application moving next to coverage of Java virtual machinesand their impact on the performance of an application. This sectionwill include a discussion of the HotSpot JVM, critical parameters thataffect the operation of the JVM, and especially garbage collectionalgorithms. We will describe how JVM garbage collection works,how it impacts the performance of an application, what parametersare available for tuning its operation and how to tune theparameters.

The next section will cover tuning tips that may be used duringdevelopment of an application, covering different areas of the Javaprogramming language such as the object lifecycle, I/Operformance, serialization, the choice of collections and parsing.

The workshop will describe some of the benchmarks that areavailable to assist you with building a high-performance Javaapplication and load-testing tools for helping you detect bottlenecks.

We will close the session by bringing all the information togetherthrough a performance tuning exercise of an actual three-tieredapplication. While profiling the sample application, we will apply aworkload and locate performance problem spots.

Odysseas Pentakalos is Chief Technology Officer of SYSNETInternational, Inc., where he focuses on providing his clientsconsulting services with performance management of computersystems and architecture, design and development of largedistributed systems that utilize Java, XML and J2EE technologies.His clients have included major government agencies andcorporations such as NASA Goddard Space Flight Center, the ArmyResearch Lab, Sun Microsystems, Concert Communications,KPMG, Bearing Point and Northrop Grumman, IT. He holds a Ph.D.in Computer Science from the University of Maryland. He haspublished dozens of papers in conference proceedings andjournals, is a frequent speaker at industry conferences and is theco-author of the book Windows 2000 Performance Guide that ispublished by O'Reilly.

John P. Pilch

Performance Capacity Solutions

Afternoon only

CPE as MIS or Information for Making

Decisions

Ever get caught in the loop of designing a report for managementand not quite getting it right? Reporting transmits data that has beenprocessed into a form that is meaningful to the recipient and is ofreal perceived value in a current or prospective decision. It ispossible that there is a disconnect between the source of theinformation and the recipient of the information. This leads to theexamination of the nature of the decision and the nature of thedecision maker.

Decisions can be broadly categorized as operational, managerial(tactical),or strategic. The scope of the decision determines thescope of the data sources that must be processed into usableinformation to support the decision. In addition, the decisioncategory indicates the characteristics of the information such astime frame, scope, frequency, and precision.

Information, per se, cannot make a decision. It can, at best, reducethe risk surrounding the decision. The approach of this workshop isto first analyze the question to be answered and the managerasking the question.The characterization of the nature of thedecision and maker leads to the characteristics of the information.The information sought leads to the potential data sources andsome indication as to the precision of the information reported.Once the data sources have been identified, they can be processedand organized to aid in decision making.

John Pilch has a long career as a computer scientist and aprofessional educator. He possesses the unique background ofover thirty years of industrial experience at world recognizedresearch and system engineering firms with over thirty yearsexperience teaching computer science at the graduate level. After atwenty year career at Bell Laboratories and Bell CommunicationsResearch, he is consulting on the deployment of large multi-tier,multi-platform applications. His experience and training includeshardware configuration sizing, performance benchmarking,application modeling, system instrumentation, and performanceanalysis on a variety of platforms.

John has consulted with firms in a variety of industries includingtelecommunications, financial, and pharmaceuticals on wideranging management issues from organizational structure, tacticaldeployment of systems, and three to five year planning.

John has a long association with CMG. He has presented numeroustutorials and papers at the National Conference and at RegionalCMG Meetings. He has served terms as National Director andTreasurer along with many years on the Conference PlanningCommittee.

17

CM

G2

00

5•

Or

la

nd

oSUNDAY WORKSHOPS

18

Jack Probst

Pink Elephant

Afternoon only

ITIL at Warp Speed

Want to get up to speed on ITIL, but don't havetime for a 2-day class? No problem. This is thetheory from Essentials, condensed into 1/2 a day!This rapid-fire, wise-cracking, no-punches-pulledrendition of ITIL Essentials, will leave yougasping for breath with cramps in your note-taking hand. The Service Support and ServiceDelivery processes will never be the same!

Key learning objectives include:• General introduction to the ITIL framework• Overview of ITIL Service Support processes• Overview of ITIL Service Delivery processes• Better understanding of IT service

management best practices

Jack Probst is an Executive Consultant with PinkElephant, a leading provider of ITIL consulting,education and conferences. Jack's is responsiblefor working with senior management of clientfirms to identify the opportunity of the ITILpractices, discuss the challenges and provideguidance for implementation. Jack also conductsadvanced training and education programs onITIL and management issues associated withprocess implementation. .

Prior to joining Pink Elephant, Jack worked forNationwide Insurance where he led the Officer ofProcess and Governance which facilitated effortsto implement consistent processes including ITIL,IT portfolio management, project developmentand implementation, technology acquisition andIT contributions to mergers and acquisitions.Jack also served as the CIO for ScottsdaleInsurance Company, a Nationwide subsidiary.Before joining Nationwide, he spent 17 years withthe Crum and Forster companies, a division of

Xerox Corp.

Jack is a CPCU, holds a MBA from Georgia StateUniversity, a bachelor's degree in mechanicalengineering from Georgia Tech and is certified asan ITIL Service Manager. Jack's hobbies includeteaching, astronomy and radio control airplanes.

Cathy A. Wright

BT

Morning only

ITIL - What's IT All About Then?

ITIL – The IT Infrastructure Library – is aninternationally recognised framework supportingbest practices in IT Service Management. Itcompliments international standards such asISO20000 (formerly BS15000), and ITGovernance programmes such as SarbanesOxley

This workshop will cover the key elements of theITIL framework relating to Service Management.It will also look briefly at some of the relatedareas touched on by the ITIL publications suchas Infrastructure Management and SoftwareAsset Management.

A further section will concentrate on IT ServiceConfiguration Management and its position at theheart of a process oriented, business alignedService Management Strategy. It will look at thediffering types of Configuration Management andtheir interfaces both to each other and to thewider Service Management process areas. Thissection will also include a discussion on some ofthe tools appearing in the emergingConfiguration Management markets.

We will also look at some of the mythssurrounding ITIL, some of the cultural andtechnical barriers to successful implementation,and what the future holds in terms of internationalstandards such as ISO20000

Sections:• ITIL Overview• IT Service Configuration Management • IT Governance & Regulatory issues• ITIL – Myths, Culture, Futures

Cathy Wright is the IT Assets and ConfigurationManager for British Telecom. Her IT experiencealso extends to the Capacity and Performancearena where she ran a team of specialistscovering mainframe and mid range capacityplanning for BT. She holds ITIL accreditation andhas presented technical and ITIL related papersat many CMG conferences in recent years. Cathywas also the Chair of UKCMG.

DE

CE

MB

ER

• 4-9

MONDAY USER GROUP MEETINGS & DESCRIPTIONS

19

CMG2005 MONDAY SCHEDULEMEETING SCHEDULE

VICE PRESIDENT’S REGIONAL,

INTERNATIONAL OFFICERS, AND

ADVISORY COUNCIL MEETING 9:00 AM - 12:00 PM

INTERNATIONAL OFFICERS MEETING 1:00 PM - 3:00 PM

SESSION MONITORS & CHAIRS MEETING 4:30 PM - 5:30 PM

FIRST-TIME ATTENDEE ORIENTATION 4:30 PM - 5:30 PM

ANNUAL BUSINESS MEETING / OPENING SESSION 6:00 PM -

PARS IMMEDIATELY FOLLOWING

BUSINESS MEETING

MONDAY USER GROUPS

BMC Software 7:30 AM - 6:00 PM

Demand Technology Software, Inc. 1:00 PM - 5:00 PM

ISM (the Information Systems Manager) 8:30 AM - 4:00 PM

MVS Solutions Inc. 8:00 AM - 5:00 PM

OPNET Technologies 9:00 AM - 5:00 PM

SAS Institute Inc. 1:00 PM - 5:00 PM

TeamQuest Corporation 1:00 PM - 5:00 PM

CM

G2

00

5•

Or

la

nd

oMONDAY USER GROUP MEETINGS & DESCRIPTIONS

20

BMC Software2101 CityWest Blvd Houston, TX 77042Phone: 713-918-8800www.bmc.com

BMC Software will hold its Annual PerformanceConference at CMG on Monday, December 5th, atGaylord Palms Resort & Convention Center. TheConference features the latest solutions forperformance, capacity management of enterprise-wideapplications as well as automated provisioning. BMCSoftware will present how our solutions ensure theperformance and availability of business-criticalapplications in today’s physical and virtualized ITenvironments. You will learn how to proactively identifysystem bottlenecks; monitor, manage and automatesystem performance, manage the capacity of existingand future systems as well as automaticallyprovisioning the right resources at the right time - rightnow! Please join us at our Annual PerformanceConference on Monday, December 5th!

Demand Technology Software, Inc.1020 8th Ave. S. #6 Naples, FL 34102Phone: 239-261-8945 Fax: 239-261-5456www.demandtech.com

Demand Technology Software, the developer ofNTSMF, is proud to sponsor its eighth annual WindowsPerformance Symposium, a half-day event devotedexclusively to a discussion of topics of interest toCMGers responsible for managing Microsoft Windows2003 Server environments. Service-level managementfor Microsoft-powered web sites, the advent ofinexpensive 64-bit Windows computing, and serverconsolidation best practices are among the topicsplanned. The full agenda for this event will not befinalized until this fall so that we can showcase thelatest technology and developments. Please check theconference Final Agenda for complete details.However, attendees can be assured, as in years past,there will be no vendor marketing fluff!

ISM (the Information Systems Manager)One Bethlehem Plaza Bethlehem, PA 18018Phone: 610-865-0300 Fax: 610-868-6277www.perfman.com

ISM’s PerfMan* performance and capacitymanagement solutions provide sophisticated, yet easy-to-use tools and services to manage even the mostcomplex IT environments more efficiently andeffectively.

Clients join ISM each year at CMG to meet other users,interact with PerfMan designers and developers, learnabout new PerfMan features, and discuss generalcapacity and performance management issues.

A general session with information for all will befollowed by dual-track sessions (z/Architecture &Windows/UNIX/Linux). These sessions will be followedby lunch.

If you’d like to attend ISM’s 2005 User Group Meeting,please pre-register at our website: http://www.perfman.com/CMGregistration

MVS Solutions Inc.8300 Woodbine Ave Markham, ON L3R9Y74th Floor CanadaPhone: 905-940-9404 Fax: 905-940-5308www.mvssol.com

You’re invited to the annual ThruPut Manager CMGUser Group session, to be held from 9:00 am to 4:30pm. We can promise you a day of ThruPut Managertechnical news, ‘how to’ sessions, an update onThruPut Manager 6 Automation Edition and itsimplementation, and a chance to meet and talk to otherThruPut Manager users. All our customers are invitedto meet with Jose Danobeitia, the chief architect ofThruPut Manager, Nancy and Martin for an interestingday of discussions on ThruPut Manager now and in thefuture. To reserve your space please contact MartinWills at 905 940 9404 or email: [email protected]

OPNET Technologies7255 Woodmont Ave. Bethesda, MD 20814Phone: 240.497.3000 Fax: 240.497.3001www.opnet.com

TBD

DE

CE

MB

ER

• 4-9

MONDAY USER GROUP MEETINGS & DESCRIPTIONS

21

SAS Institute Inc.100 SAS Campus Drive Cary, NC 27513Phone: 919-677-8000 Fax: 919-677-4444www.sas.com

SAS invites you to our CMG Monday Users Group onDec. 5 to hear how SAS customers use SAS (r) ITManagement Solutions to address their IT challenges.In turn, SAS will unveil SAS IT Resource Management3.1, the next generation of IT intelligence. Join us todiscover how SAS IT Management Solutions allow youto align IT delivery with business demands to optimizeoverall profitability and competitiveness through:

• Sophisticated analytical reporting and data visualization.

• Reliable information on IT usage and costs.• Integrated and intuitive products for IT management

across the enterprise.

TeamQuest CorporationOne TeamQuest Way Clear Lake, IA 50428Phone: 641-357-2700 Fax: 641-357-2778www.teamquest.com

Join TeamQuest at its annual CMG Users Meeting onMonday December 5 from 1 to 5PM. Customers andIT professionals interested in performancemanagement, capacity planning, and IT ServiceOptimization are invited to the following presentations:

TeamQuest technical staff and customers addressingspecific performance management and capacityplanning issues.

Product managers and engineers discussing the latestrelease of TeamQuest Performance Software and apreview of upcoming features and functionality.

Customer presentations featuring real world examplesof how significant dollars have been saved usingcapacity planning to consolidate servers and predictresource requirements.

DE

CE

MB

ER

• 4-9


23


• TUESDAY SESSIONS

• WEDNESDAY SESSIONS

• THURSDAY SESSIONS

• FRIDAY SESSIONS

TUESDAY, DECEMBER 6TH AT-A-GLANCE

24

TIME SUBJECT SESSION AUTHOR TITLE

8:00 AM Core 301 K. Mani Chandy Sense and Respond Systems

9:15 AM Core 311 Dr. Lloyd G. Williams QSEM: Quantitative Scalability Evaluation Method

9:15 AM zOS 312 Dr. Annie W. Shum A Multi-Tiered Approach With Data Normalization To Analyzing CPU Metrics

9:15 AM ITIL 313 Chris Molloy ITIL Capacity Management Deep Dive

9:15 AM Hot 314 Daniel A. Menasce Virtualization: Concepts, Applications, and Performance Modeling

9:15 AM Core 315 Peter Johnson Scaling up the JBoss Application Server

9:15 AM Core 316 TBD Are Integrated Management solutions the recipe for an invisible and secure IT infrastructure?

9:15 AM Core 317 Dr. James Bouhana Analysis of Workload Alerts in Consolidated Servers

10:30 AM *nix 321 Mark M. Maccabee Testing Scalability of a WebLogic Application

10:30 AM *nix 321 William R. Sullivan Overcoming Limitations to Java Application Scalability

10:30 AM zOS 322 Kenneth D. Williams MVS Application Performance Management

10:30 AM Core 322 TBD UKCMG Best Paper:

10:30 AM Stor 323 Bruce Naegel SMI (SNIA) Performance Update

10:30 AM Core 324 Gene P. Fernando To V or not to V: A Practical Guide to Virtualization

10:30 AM Hot 325 Dr. Toufic Boubez Policy Driven SOA

10:30 AM Core 326 Glenn O’Donnell Making Sense of the Performance Riddle

10:30 AM ITIL 327 Avtar Dhillon Achieve IT Agility by Integrating SOA with ITIL Based BSM

10:30 AM Core 327 Dr. Thomas E. Bell Determining Architectures of Existing Systems

2:00 PM Core 331 Henry H. Liu Service Demand Models for Enterprise Software Applications

2:00 PM zOS 332 Craig Hodgins Java Performance on z/OS: A Report from the Front Lines

2:00 PM Stor 333 Mel Boksenbaum Storage Performance Council Panel Discussion

2:00 PM Hot 334 Peter J. Weilnau Measuring Up for Server Virtualization

2:00 PM Core 335 Shanti Subramanyam Performance Management of a J2EE Application to Meet Service Level Agreements

2:00 PM Hot 336 Herb VanHook Opportunities and Challenges of the SOA World

2:00 PM Core 337 Michael Wiener Calculating Expected Reliability of Systems and Hardware

3:45 PM Core 341 Michael D. Maddox Using Fuzzy Logic to Automate Performance Analyses

3:45 PM zOS 342 Don Deese Introduction to zSeries Application Assist Processor (zAAP)

3:45 PM Stor 343 Bruce McNutt Disk Arm Management of Competing Workloads

3:45 PM Core 344 Prof. Ethan Bolker Virtual Performance Won’t Do: Capacity Planning for Virtual Systems

3:45 PM Core 345 Peter Johnson Java Garbage Collection Performance Analysis 201

3:45 PM Hot 346 Eric Pulier Learning to Play Nicely

3:45 PM Hot 347 Chris Molloy Introduction to Data Center Markup Language (DCML)




25


8:00 AM Win 401 Amy D. Wohl The China Era

9:15 AM Hot 411 Dr. Jeffrey P. Buzen SOA and the Social Order - City Planning, Post Office & Business Protocols

9:15 AM zOS 412 Robert R. Rogers How Do You Do What You Do When You’re a CPU?

9:15 AM Core 413 Peter Sevcik The Application Performance Index (Apdex) Standard+

9:15 AM Stor 414 Charles T. McGavin Jr. Smoke and Mirrors – A Survey of Remote Replication Technologies

9:15 AM Hot 415 Nick Gall Reflection: The Next Big Thing After SOA

9:15 AM Core 416 Alexander Podelko Workload Generation: Does One Approach Fit All?

9:15 AM Core 417 Dick Arnold Reactive Capacity Planning – An Alternative

10:30 AM *nix 421 Jaqui Lynch Linux Performance Tuning

10:30 AM zOS 422 Glenn R. Anderson WebSphere Application Server for z/OS Version 6 - Performance Tuning

10:30 AM Win 423 Mark B. Friedman Virtual Memory Constraints in 32-bit Windows: an Update

10:30 AM Stor 424 Michael A. Salsburg A Management Framework For Petabyte-Scale Disk Storage

10:30 AM Stor 424 Kathleen N. Hodge Database Disk to Disk Backups Using ATA Disk

10:30 AM Net 425 Neil Carter Intrusion Detection/Prevention Devices - Are They Protecting Your Network -or Hampering It?

10:30 AM Core 425 Jack B. Woolley Want to Know WHY Response Time is So Long? Listen to the Wire.

10:30 AM Core 426 G Jay Lipovich Is It Time for Capacity Planners to Hang Up Their Cleats?

10:30 AM Core 427 Kevin Mobley Using Six Sigma to Define the Focus of Software Performance Engineering

2:00 PM Core 431 Dr. Jeffrey P. Buzen Workload Characterization for Parallel Processing Environments

2:00 PM zOS 432 Tony Ruberry zSeries Capacity Management - a True Story

2:00 PM Win 433 Kevin Kline Performance Baselining, Benchmarking and Monitoring for Microsoft SQL Server

2:00 PM Stor 434 Dan W. Yee I/O Performance Characteristics for Volume Managers on Linux 2.6 Servers

2:00 PM Hot 435 Ken Traub Radio-Frequency Identification at Enterprise Scale

2:00 PM Core 436 TBD CMG Australia Best Paper:

2:00 PM Core 437 Irving Smith Wholesale Distributed Capacity Planning

3:45 PM Core 441 Dr. Jozo J. Dujmovic QNS - an Online System for the Study of Queuing Models

3:45 PM zOS 442 Ian C. Baldwin Migrating to z-990 - A User Experience

3:45 PM Win 443 Michael A. Salsburg, PhD Panel: IT’s Role in Business Performance Management

3:45 PM Stor 444 Robert Rogers Information Classification and Service Level Objectives for Information Lifecycle Management

3:45 PM Net 445 Prof. James Westall Performance Tuning of Gigabit Network Interfaces

3:45 PM Hot 446 Dr. Annie W. Shum Through the Prism of Fractals: Why SOA Should Reflect the Natural Order

3:45 PM Core 447 Chandra Lanka Capacity Planning for Shared Middleware Environments; A Methodology

WEDNESDAY, DECEMBER 7TH AT-A-GLANCE

26


8:00 AM zOS 501 Cheryl Watson z990 Performance and Capacity Planning Issues

8:00 AM zOS 502 Stephen R. Guendert Taking FICON to the Next Level-Cascaded High Performance FICON

8:00 AM Win 503 Jeffry A. Schwartz Understanding and Interpreting SQL Server Performance Counters

8:00 AM Hot 504 Cathy Nolan Encryption Primer: An Introduction to Data Protection

8:00 AM Net 505 Bernie Davidovics Measuring End User Response Time - With a Passive Network Probe

8:00 AM Core 506 Dr. Pierre M. Fiorini Software Performance Engineering Considerations in Unreliable Computing Environments

8:00 AM Core 507 Denise P. Kalm Survivor - The Corporate Jungle

9:15 AM zOS 511 Ned A. Diehl DB2 CPU and Response Metrics

9:15 AM Core 512 Ellen M. Friedman Measuring Performance in the Lab and Validating it in Production

9:15 AM Hot 513 Marina Cismas Redefining Capacity Planning for Grid Computing

9:15 AM Core 514 Linwood Merritt Closing the Gaps – Understanding Capacity Summarization

9:15 AM Net 515 Laura Knapp Introduction to Simple Network Management Protocol (SNMP) Version 3

9:15 AM Hot 516 Dr. Connie U. Smith A Performance Model Web Service

9:15 AM Core 517 James A. Yaple Benchmarking 101

10:30 AM zOS 521 Kathy Walsh The Mechanics of Developing a High Quality Capacity Plan

10:30 AM zOS 522 Thomas A. Halinski Unveiling of DB2’s DDF: SQL Revealed via the Gestalt Perspective

10:30 AM *nix 523 Richard K. Roehl Designing a Fairer Round Robin Scheduling Algorithm

10:30 AM *nix 523 Serge Tessier Capacity Planning for UNIX System Metrics

10:30 AM Stor 524 Stephen R. Guendert Proper Sizing and Modeling of ESCON to FICON Migrations

10:30 AM Net 525 Nalini J. Elkins One-Minute TCP Stack Analysis

10:30 AM Core 526 Dr. Mary R. Hesselgrave Underground SPE: Moving from Performance QA to SPE

10:30 AM Core 526 Heidi A. Gilmore Choosing a Load Testing Strategy

10:30 AM *nix 527 Robert F. Patterson Capturing System Data Using Native Commands

10:30 AM *nix 527 Igor A. Trubin Capturing Workload Pathology by Statistical Exception Detection System

2:00 PM zOS 531 Peter Enrico WLM Sysplex Management

2:00 PM *nix 532 Rick Lebsack Facilitated Discussion: Future of the Performance Field 2005

2:00 PM *nix 533 James Holtman Visualization of Performance Data

2:00 PM Stor 534 Prof. Robert Geist Enhancing Web Server Performance Through the Use of a Drop-In, Statically Optimal Disk

Scheduler

2:00 PM Net 535 Chris Loosely Measuring Actual and Perceived User Experience on the Web

2:00 PM Hot 536 Richard Soley The Model-Driven (R)evolution

2:00 PM zOS 537 Michael P. Swanson 8 Great Myths of Software Asset Management

3:45 PM zOS 541 Ivan L. Gelb Panel: zSeries Performance Q & A

3:45 PM Core 542 TBD CMG Italia Best Paper:

3:45 PM Hot 543 David R. Morley CMG: The Early Years

3:45 PM Stor 544 Chao Li Analytic Way for Performance Management of SAN

3:45 PM Net 545 Dr. Swami Ramany Workload modeling of Stateful Protocols Using HMMs

3:45 PM Hot 546 William L. Shelden, Jr., Ph.D. Modeling VMware ESX Performance

3:45 PM Core 547 Dr. Anatoliy Rikun Using Principal Component Method for Performance Data Compression and Analysis

THURSDAY, DECEMBER 8TH AT-A-GLANCE




27


8:00 AM zOS 603 Ivan L. Gelb WLM Caused Pain and Pleasure

8:00 AM zOS 604 Dr. Bernard Domanski Speaking SOA and Web Services: .NET and the Mainframe

8:00 AM Net 605 Bill Furnas Managing J2EE Applications with ARM 4.0 and a Tutorial of the Open-Source ARM 4.0 Java

and C SDK

8:00 AM Core 606 Stuart Plotkin Distributed Resource Reclamation: Enterprise Shared Servers

9:15 AM Win 613 Chris Greco Monitoring the Monitors (If a Tree Falls in a Forest, Will it Affect the Network Bandwidth)

9:15 AM zOS 614 Robert E. Chaney DDF Performance Analysis - Does it Really Have to be This Complicated?

9:15 AM Net 615 Jie Lu Performance Modeling and Analysis of Web Switch

9:15 AM Core 616 Dr. Charles A. Letner A Methodology for Predicting the Scalability of Distributed Production Systems

10:30 AM Core 623 Frank Bereznay Making Your Web Portal a Dynamic Website

10:30 AM zOS 624 Charles E. Hackett LSPR Benchmark Converter

10:30 AM Core 626 Rupa S. Joshi It’s All About Statistics!

FRIDAY, DECEMBER 9TH AT-A-GLANCE

DE

CE

MB

ER

• 4-9

SESSION DESCRIPTIONS

29

SESSION DESCRIPTIONS

CM

G2

00

5•

Or

la

nd

oTUESDAY SESSIONS

SESSION 301 TUESDAY 8:00 AM - 9:00 AM

PAPER 5401 NON-SPECIFICFUNDAMENTALS/CORE COMPETENCY No Paper

Sense and Respond SystemsK. Mani Chandy, California Institute of Technology

Sense and respond systems sense, and then respond, to opportunitiesand threats. A sense and respond platform can be configured todevelop specific sense and respond applications. Sense and respondapplications arise in homeland security, healthcare, finance, supplychains, energy, environmental protection, security and - mostimportantly for this talk - the management of IT infrastructure. TheCMG community can make a profound impact on the space of senseand respond applications because this community has expertise in therelevant technologies and mathematics: CMG papers deal withmeasurement of asynchronous events, statistics, probabilistic models,information fusion, and real time. Sense and respond applications areincreasingly important to society, and CMG has a major role to play.This talk will survey the field of sense and respond applications;identify applications spaces that are critical for society; survey thetechnologies and software architectures used in these applications;describe fundamental problems in experimentation, systems designand theory; explore the rate of growth of this space; show how theCMG community’s experience is directly relevant; and discussexperience with developing applications from platforms.

Chair: Catherine S. Nolan, Bank of America


PAPER 5165 HOW-TOFUNDAMENTALS/CORE COMPETENCY Z/OS

QSEM: Quantitative Scalability

Evaluation MethodDr. Lloyd G. Williams, PerfXConnie U. Smith, Performance Engineering Services

While scalability is important to today’s software applications, feworganizations understand how to quantitatively evaluate theirsoftware’s scalability. This paper describes the Quantitative ScalabilityEvaluation Method, QSEM. QSEM uses straightforwardmeasurements to quantify the scalability of a software application. Theresults provide an understanding of the application’s scalability thatmakes it possible to extrapolate behavior to larger configurations withconfidence. The seven steps of the QSEM method are described andillustrated with a case study.

Chair: Ellen M. Friedman, SRM Associates, Ltd.


PAPER 5159 INTRODUCTORYMAINFRAME (Z/OS) Z/OS

A Multi-Tiered Approach With Data

Normalization To Analyzing CPU

MetricsDr. Annie W. Shum Boris Ginis, BMC Software

In IT management, CPU utilization is arguably the most basicperformance metric. It is also one of the most multi-faceted, with IBM’sRMF and SMF providing over 60 variations of zSeries CPU utilizations,spanning differing degrees of granularity and complexity. So whichCPU utilization metrics are right for you? Why, when, and how do youuse each of them? We advocate a new, multi-tiered approach built ona broad-based data normalization technique. In addition, we also offerguidance on how to select CPU utilization metrics to address specificproblems, such as capacity planning.

Chair: Shana J. Bereznay, ACS


PAPER 5088 ADVANCEDITIL NON-SPECIFIC

ITIL Capacity Management Deep DiveChris Molloy, IBM

ITIL is continuing to grow in acceptance in IT environments as a modelfor best practices. This paper provides a low level analysis of the ITILcapacity management discipline. The paper describes the differencesbetween business, service, and resource capacity management, andthe need for each to have a proactive capacity management process.The paper will describe the elements needed for an ITIL-basedcapacity plan, a discussion on what several companies have done withITIL capacity management, and lessons learned from implementingITIL capacity management discipline in their environment.

Chair: Minda E. Larson, Storage Technology Corporation


PAPER 5189 TECHNICALHOT TOPICS NON-SPECIFIC

Virtualization: Concepts,

Applications, and Performance

ModelingDaniel A. Menasce, George Mason University

Virtualization was invented more than thirty years ago to allowexpensive mainframes to be shared among different applications. Ashardware prices went down, the need for virtualization faded away.More recently, virtualization at all levels (system, storage, and network)became important again to improve system security, reliability, and toreduce costs. This paper explains the basics of system virtualizationand addresses performance issues related to modeling virtualizedsystems using analytic performance models. A case study on serverconsolidation is used to illustrate the points.

30

SESSION DESCRIPTION LEGEND

SESSION NUMBER DAY START TIME - END TIME

PAPER NUMBER SESSION TYPESUBJECT AREA PLATFORM

No Paper *

TitleAuthor, CompanyCo-Author, CompanyCo-Author, Company

Abstract....

Chair:

* Session Handouts may be available on the CMG websitefollowing the conference: www.cmg.org/membersonly

DE

CE

MB

ER

• 4-9

TUESDAY SESSIONS

31


PAPER 5009 TECHNICALFUNDAMENTALS/CORE COMPETENCY WINDOWS NT/XP/ETC.

Scaling up the JBoss Application

ServerPeter Johnson, Unisys

This paper presents our experience with tuning the JBoss ApplicationServer to run the SPECjAppServer2002 benchmark. We used ourknowledge of high-performance, highly scalable systems to push theJBoss Application Server to its limits. This paper describes how thoseperformance gains were accomplished, highlights various significantimprovements, and provides information about the performance gainsthat were attained. The reader should be able to use the informationpresented in this paper to tune his or her own applications deployed tothe JBoss Application Server.

Chair: Kadhar Masthan, Cognizant Technology Solutions


PAPER 5418 No PaperFUNDAMENTALS/CORE COMPETENCY

Are Integrated Management Solutions

the Recipe for an Invisible and

Secure IT Infrastructure? TBD, Computer Associates

As businesses today are demanding greater flexibility and efficiency,we see that the IT Organizations are demanding the managementsoftware providers to deliver a complete, integrated and open solutionfor policy-driven, adaptive Services Oriented Infrastructure (SOI) thatcan expose assets as composite IT services and enable a contextuallyrelevant view of business processes. In this presentation, we willdiscuss how we could optimize the delivery of business servicesthrough an invisible and secure infrastructure that can transparentlyadapt to changes in business conditions. We will also highlight howManagement solutions that are integrated across the IT stack arebetter suited to achieve the vision of an invisible and secureinfrastructure.

Chair: Chris L. Molloy, IBM


PAPER 5024 TECHNICALFUNDAMENTALS/CORE COMPETENCY NON-SPECIFIC

Analysis of Workload Alerts in

Consolidated ServersDr. James Bouhana, Performance International, Inc.Mike Tsykin, Fujitsu Australia Limited

Traditional alerting on performance metrics works well when serversare dedicated to one application — the likely source of the alert isknown. However when multiple workloads are hosted on aconsolidated server, probable cause analysis for alerts becomes morechallenging. An approach is presented for stratifying alerts by bothmetric and workload so that the distribution of alerts across workloadsfor each performance metric can be seen. An Alerts Map gives a birds-eye view of all alerting activity. The challenges and successes inimplementing the multi-workload analysis are also discussed.

Chair: Susan Schreitmueller, IBM

SESSION 321 TUESDAY 10:30 AM - 12:00 PM

PAPER 5175 TECHNICALUNIX/LINUX UNIX/LINUX

Testing Scalability of a WebLogic

ApplicationMark M. Maccabee, IBM

To improve the scalability of an existing production application it wasrewritten. The rewritten application was based on a new architecture,using J2EE technology (WebLogic product). The new Application wasstructured with multiple sub-applications using a set of commoncomponents. The common component expected to be the most heavilyused was identified; this was the Access Control Facility (ACF).Development of ACF was accelerated to allow testing, analysis andimprovements to occur well ahead of the development of the othercomponents. This paper describes the scalability testing of ACF.

Chair: Elizabeth A. Somers, IBM Corporation

PAPER 5166 ADVANCEDUNIX/LINUX Z/OS

Overcoming Limitations to Java

Application ScalabilityWilliam R. Sullivan, WHAM Engineering & Software

Java is a robust language as well as development platform, for bothserver and client applications. The JVM is the undergirding supportstructure for both the language and the runtime environment. We lookat the limitations which the JVM places upon applications from both aCPU usage and a serialization perspective. Several real examplesfrom previous projects are examined as test cases, and the solution tothe limitations is provided.

Chair: Elizabeth A. Somers, IBM Corporation



MVS Application Performance

ManagementKenneth D. Williams, CPT Global

It is quite possible, and sometimes very easy, to make dramaticsavings in the load on the system caused by some applications. Thispresentation explains the reasons behind application performancetuning, the tool kit required, goes into some case studies or war stories,and finally gives a set of performance hints and tips to reduce overallCPU and elapsed time in applications.

Chair: Christopher Mellgren, Fidelity Investments

PAPER 5422

UKCMG Best Paper:TBD

TBD

Chair: Christopher Mellgren, Fidelity Investments

BUSINESS MANAGEMENT SHOWCASE

CM

G2

00

5•

Or

la

nd

oTUESDAY SESSIONS


PAPER 5137 PANELSTORAGE UNIX/LINUX

No Paper

SMI (SNIA) Performance UpdateBruce Naegel, VERITASBill Zahavi, EMC CorporationJim Nagler, VERITAS Software

The first release of the SMI Performance specification as part of the 1.1specification will be available this year. The SMI Performance group isworking on the next iteration of this specification, extending it to theentire storage network as well as methods to tune the disk array’sperformance. This session will describe the current status of thisperformance sub profile including:

• Use cases to be addressed with the sub-profile• Proposed measurement methods and parameters• Status to date in building prototypes of this functionality.

Chair: Gary Y. Furukawa, Southern California Edison


PAPER 5014 INTRODUCTORYFUNDAMENTALS/CORE COMPETENCY WINDOWS NT/XP/ETC.

To V or not to V: A Practical Guide to

VirtualizationGene P. Fernando, BMC Software

Virtualization has become so widespread that industry pundits havecalled it a megatrend. Many forms of Virtualization are available today,each with its own set of potential gains. Of course, with every newtechnology comes a new set of problems. This paper explores thebenefits of Virtualization and discusses the difficulties in measuring theresults in the real world. A case study is included to present a practicalapproach for measuring and forecasting growth for virtual servers.

Chair: Michael Gardner, RBC Financial Group


PAPER 5407 No PaperHOT TOPICS

Policy Driven SOADr. Toufic Boubez, Layer 7 Technologies

The original goal of the Service Oriented Architecture (SOA) concept,and its implementation technologies such as Web services, was tobuild flexible, loosely coupled systems. But any two components in asystem that communicate with each other are by definition coupled toa certain extent. The goal therefore is to lessen that coupling, or, to putit another way, to loosely couple components in systems. That meansremoving or lessening the runtime dependencies between them. Thebest mechanism to achieve that is to delegate as much as possible theruntime tasks to the infrastructure. In order for this to work, contracts,requirement and capabilities need to be defined and automatedthrough a declarative, configurable, and manageable mechanism.WSDL is the contract language for Web services. But WSDL is far frombeing adequate as a contract language for SOA. The required level ofabstraction for SOA sits at the Policy level. Policies contain assertionsabout the operational interfaces for components in an SOA. Theseinclude credential preferences, authentication and authorizationmechanisms, signature and encryption preferences, identity sources,routing, transformations, versioning, reliable messaging and others.This talk will introduce the concept of Policy Driven SOA and discussPolicy as the new contract abstraction for SOA.


PAPER 5416 No PaperFUNDAMENTALS/CORE COMPETENCY

Making Sense of the Performance

RiddleGlenn O’Donnell, EMC

Service performance is a “riddle wrapped in a mystery inside anenigma” to the extreme majority of IT organizations. To trulyunderstand this riddle, one must navigate a complex labyrinth ofinterrelated technology and business components. New compositeapplications and distributed services exacerbate this scenario. Nohuman is capable of comprehending all of this complexity, soautomation technology, process, and standardization are the keys tocontrolling our environments and to leveraging this complexity forcompetitive benefit. Some early successes prove it is possible, albeitdifficult, as cultural shifts never come easily. Beware of grand promisesof automated simplicity. Such nirvana remains many years from reality,if ever.

Chair: David C. Thorn, SunGard Availability Services


PAPER 5195 INTRODUCTORYITIL NON-SPECIFIC

Achieve IT Agility by Integrating SOA

with ITIL Based BSMAvtar Dhillon, BMCAnnie W. Shum

Business Service Management (BSM), rooted in ITIL, helpsorganizations manage IT from a business perspective. By identifyingand mapping business-critical processes to the underlying ITinfrastructure and services, BSM connects key business services tothe IT services that manage them, such as routers, servers, andapplications. While BSM does the mapping, SOA serves as theframework that connects the infrastructure. Hence, the BSMmethodology and SOA are synergistic and the integration of BSM intoSOA is pivotal to IT agility.

PAPER 5185 HOW-TOFUNDAMENTALS/CORE COMPETENCY NON-SPECIFIC

Determining Architectures of Existing

SystemsDr. Thomas E. Bell, Rivendel Consultants

An architectural description of an existing application is often needed,especially for performance work. The architecture is the highest leveldescription of an implemented system in response to its requirements;it is needed to understand interactions for performance analysis.Academic descriptions can help, but existing system documentation,combined with extensive interviews, are needed to discover the mostimportant characteristics.

Chair: Nancy A. Sabatini, Bank of America

32


DE

CE

MB

ER

• 4-9

TUESDAY SESSIONS

33

SESSION 331 TUESDAY 2:00 PM - 3:00 PM


Service Demand Models for

Enterprise Software ApplicationsHenry H. Liu, DST Innovis, Inc.

Queuing models are solved with process probability distributions forsystem performance analysis and capacity planning. However, themodels depend on service demand as one of the input variables whichcan be obtained only through measurements of an existing system.This kind of dependency makes it difficult to project the performanceand capacity requirement of an enterprise software application prior todeployment in production. In this paper, we demonstrate that thisconstraint can be removed. Our models help support the notion of“predict and build” for developing enterprise software applications.

Chair: Frank M. Bereznay, Kaiser Permanente



Java Performance on z/OS: A Report

from the Front LinesCraig Hodgins, Compuware

Java is ten years old this year and is being used more and more as thelanguage of choice for mainframe applications under WebSphere. ButJava is not COBOL. It is important that analysts understand theperformance implications of this new world. This paper will use actualfront line application examples to illustrate some of the performanceissues involved with Java running under WebSphere on the mainframe.

Chair: Warren Hayward, The TJX Companies Inc


PAPER 5149 PANELSTORAGE NON-SPECIFIC

No Paper

Storage Performance Council Panel

DiscussionMel Boksenbaum, Hitachi Data Systems

The first cross-vendor team of storage performance experts has builtthe industry’s first benchmark for storage that has become the standardfor decision making. The Storage Performance Council has sought areal-world workload to become the first benchmark that is vendor-neutral, platform independent, network storage capable. More thanthree dozen SPC-1 results have been published to date. This panelsession will discuss the status of the Storage Performance Council(SPC) and the Storage Industry standard performance benchmarksavailable and under development.


PAPER 5054 INTRODUCTORYHOT TOPICS NON-SPECIFIC

Measuring Up for Server

VirtualizationPeter J. Weilnau, ISM

Server virtualization is one of the hottest topics for the capacity plannerin 2005. How does a hardware virtualization environment like VMwareESX impact the performance measurements we have beenaccustomed to getting from Windows and Linux? If you can’t measureit, you can’t manage it, so how are we to measure and understand thisnew environment? This paper presents a lab-based study designedspecifically to aid in understanding performance measurementsavailable directly from VMware ESX. The study sheds light on theimpact of virtualization on Windows and Linux performanceinformation.

Chair: Claire S. Cates, SAS


PAPER 5129 TECHNICALFUNDAMENTALS/CORE COMPETENCY UNIX/LINUX

Performance Management of a J2EE

Application to Meet Service Level

AgreementsShanti Subramanyam, Sun Microsystems Inc.

As Service Level Management increasingly encompasses multi-tierJ2EE applications, it is no longer sufficient to do one-time capacityplanning or static performance analysis and predictions. UtilityComputing requires dynamic performance management capabilitiesthat include modeling, analysis and predictions based on dynamic data.This paper analyzes various approaches to this problem and describesa methodology that can help solve this problem for real-world, multi-tierenterprise applications.

Chair: Donna S. Folkerts, IBM Global Services


PAPER 5417 No Paper

Opportunities and Challenges of the

SOA WorldHerb VanHook, BMC Software

The loosely-coupled application architectures emerging as thepreferred new wave of integration enable flexible and adaptablebusiness models, but come with their own challenges when mapped tothe traditional management and operational process disciplines. Thepromise of Service-Oriented Architectures presents alignmentopportunities between technology organizations and their businesscounterparts, as a closer mapping between real-world businessservices and application components is realized. As companies moveto a “flow computing” paradigm with the adoption of Web services,many of the lessons learned over the years in running high-performance technology environments will carry forward, but many ofthem will have to be adapted to the realities of these new architectures.This session will address the imperatives that Service-OrientedArchitecture “run-time” brings to the forefront.

Chair: Jonathan Gladstone, Bank of Montreal


CM

G2

00

5•

Or

la

nd

oTUESDAY SESSIONS

34



Calculating Expected Reliability of

Systems and HardwareMichael Wiener, Bank of America

The need to calculate the reliability and expected up-time of newsystems has become critical. This paper describes some simplemethods that can be used to determine the reliability of a currentsystem and the underlying components. It also shows how thesetechniques can be used to determine if an improvement is warrantedand if so where. The paper starts by reviewing the concepts behind thecalculations, using single system failures. It then goes onto calculatingthe probability of failure across numerous components and then showshow these techniques can be used to analyze and improve thereliability of a complex system.

Chair: Margaret R. Greenberg, Grunberg Haus



Using Fuzzy Logic to Automate

Performance AnalysesMichael D. Maddox, MCI

Fuzzy logic, also known as approximate reasoning, is an artificialintelligence technology originally conceived in the 1960s by LotfiZadeh. It has been applied successfully in many areas, includingcontrol systems. This paper defines key concepts of fuzzy logic, showsexamples of fuzzy logic in life and in real-world applications, showshow fuzzy logic can be applied to computer performance work tosimplify and speed analysis and reporting, and shows a simpleexample of program code which implements fuzzy logic. Examples arebased on MS-Windows(R), but the principles can be applied to otherOSs.

Chair: James Holtman, Convergys



Introduction to zSeries Application

Assist Processor (zAAP)Don Deese, Computer Management Sciences, Inc.

Starting with z/OS V1R6 on z890 and z990 servers, Java applicationscan run on a new type of processor called the eServer zSeriesApplication Assist Processor (zAAP). The zAAP is a relativelyinexpensive solution for installations running a large amount of Javawork. This paper presents an overview of zAAP processors, describeshow zAAP processors interact with z/OS, discusses someperformance considerations when implementing zAAP processors,and describes the data available in RMF that can be used to analyzezAAP performance.

Chair: Rupa S. Joshi, Bank of America


PAPER 5130 TECHNICALSTORAGE NON-SPECIFIC

Disk Arm Management of Competing

WorkloadsBruce McNutt, IBM

Can scheduled nighttime processing somehow escape the tyranny ofthe ‘’batch window’’? Can transactions and large-scale queries be runagainst the same database at the same time, while maintainingacceptable levels of performance? To those managing database orother systems that require access by multiple applications to acommon pool of data, such questions tend to be a key focus.Encouraged by recent developments in the SCSI standard, this paperconsiders the possibility of delegating such performance managementto the individual disk drive.

Chair: Paul Billick, Harleysville Insurance



Virtual Performance Won’t Do:

Capacity Planning for Virtual

SystemsProf. Ethan Bolker, BMC Software, IncYiping Ding, BMC SoftwareKen Hu, BMC Software, Inc

The history of computing is a history of virtualization. Each increase inthe number of abstraction layers separating the end user from thehardware makes life easier for the user, but harder for the planner whomust choose a configuration to guarantee performance.

We review the architectures of several contemporary virtual systems,and report on experiments that show how naive interpretations oftraditional metrics like “utilization” can lead planners astray. Then wepropose a generic set of performance metrics and simple predictionguidelines that can help planners manage those systems.

Chair: Tom Moulder, TREX Associates Inc.



Java Garbage Collection

Performance Analysis 201Peter Johnson, Unisys

The HotSpot® Virtual Machine for the Java 2 Platform provides avariety of garbage collection algorithms geared towards differentapplication behavior or requirements. This paper discusses the prosand cons of each algorithm, and shows how to gather and analyzestatistics provided by each algorithm. With a greater understanding ofthe garbage collection options available, and how to analyze theirperformance, the reader will be better equipped to choose the properalgorithm for use with his or her Java applications.

Chair: Charles Padamadan, ACS

DE

CE

MB

ER

• 4-9

Hotel & REGISTRATION FORMS

35

The Gaylord Palms Resort and Convention Center in Orlando:

Where CMG2005 Will Shine

CMG2005 will be held at The Gaylord Palms Resort and Convention Center, whichholds just over 1400 guest rooms and suites with beautiful gardens under its 4.5-acreglass atrium, where the weather is a perfect 72 degrees every day. Along with variousrestaurants, cafes and swimming pools are the incomparable Canyon Ranch SpaCluband nearby golf. This beautiful resort offers guests a variety of shops, lots ofentertainment and they even have alligators in the Everglades! And for those of youbringing family along, the hotel is located within minutes from the Walt Disney WorldResort and a host of other Orlando attractions.

The convention facilities are state-of-the-art, and CMG’s room block is located nearby,allowing CMGers to move between their rooms and the sessions quickly and easily.CMG will be delivering wireless connectivity in every session room, and each hotelroom offers in-room computing systems with high-speed Internet access and a host ofhotel services.

From the conference to the hotel, we are sure your CMG2005 experience will bememorable. Please come to take advantage of the CMG learning experience in a hotelyou will not soon forget.

IMPORTANT POINTS• Contact the Gaylord Palms

Resort directly by phone, or fax this completed form to the hotel to reserve your room.

• Print or type clearly to ensure correct processing.

• Use one form per room reservation.

• All reservations will require a one (1) night advance deposit per room, plus tax to guarantee room. Reservations must be guaranteed with a credit card or advanced deposit.

• To reserve your room with a cash deposit, mail a check or money order for the first night’s room and tax along with this form directly to the Gaylord Palms Resort & Convention Center for the first night’s room & tax rate.

• All cancellations must be made 72 hours prior to your check-in or you will be charged one night’s room and tax.

• Check-in time is 3:00 pm & Check-out time is 11:00 am

Gaylord Palms Resort and Convention Center

6000 W. Osceola ParkwayKissimmee, FL 34746 USA

Tel: 407-586-0000

Or: 407-586-2000

Fax: 407-586-2259

www.gaylordpalms.com

For CMG Information, call 1-800-4 FOR CMG

HOTEL RESERVATION FORM December 4 - 9, 2005 in Kissimmee, FL

Convention Group Code: A-COMPU

It’s time to start planning for CMG2005! Complete this form and mail or fax it tothe Gaylord Palms Resort and Convention Center. The CMG hotel discounted rates below apply. Be sure to mention CMG when you register! To guarantee your roomreservation at the CMG rate you must reserve your room before Tuesday,November 1, 2005 at 5:00PM Eastern Pacific Time.

Mail or fax this Hotel Registration Form to the hotel.

Please type or print clearly and make a copy for your records.

First Name:__________________ M.I.:___ Last Name:__________________

Title:__________________________________________________________

Company:______________________________________________________

Address:_______________________________________________________

City:______________________ State/Province:____ Zip/Postal Code:_____

Country:___________________ E-Mail:______________________________

Phone:__________________________ Fax:__________________________ For International numbers, please include country code.

Number of persons in party:___________ Number of rooms needed:_____________

Attendee Name:_____________________ Arriving:________ Departing:_________

Sharing With:_______________________ Arriving:________ Departing:_________

Sharing With:_______________________ Arriving:________ Departing:_________

CREDIT CARD INFORMATION � Visa � MasterCard � American Express � Other______________

Acct. #:____________________________________________ Exp. Date:___________

Name of Cardholder:____________________________________________________(Please print as it appears on the credit card)

Signature of Cardholder:___________________________________________________(Signifies authorization to charge credit card account)

Room Rates:� Single / � Double $169 + 13% tax + $10 resort fee� Triple $189 + 13% tax + $10 resort fee � Quad $209 + 13% tax + $10 resort fee� Emerald Bay add $40 to the above rates � Government Rate (Proper identification required. If calling the

hotel directly, please ask for Government rate block)

Preferences:� King Size Bed � Non-Smoking Room � Double Queen Beds

36

Choose one option from 1-7. Option 8 may be an additional option.

December 4 - 9, 2005Gaylord Palms

Orlando, FL

REGISTRATION FORM3 Easy Ways to Register!

1. Internet:www.cmg.org/conference

2. Fax:856-401-1708

3. Mail:Computer Measurement Group, Inc.P.O. Box 8500-5545Philadelphia, PA 19178-8500

For more information call 1-800-4 FOR CMG or 856-401-1700

Membership Number Important: Please provide your Membership Number. New members will be assigned a number byCMG. If unsure, please leave blank.

Please Note: Your badge will read exactly as indicated on this form. Please print or type all information clearly. Keep a copy of the completed form for your records

Name:

Job Title:

Company:

Address:

City: State: Zip:

Country:

Phone: Fax:

For Corporate Badges (options 6 & 7) indicate the person to receive CMG literature

Please include mail code

E-mail address:

Your Badge Name:

Guest Name:

Guest Badge Name:

Corporate Badges (options 6 & 7) will list company name as badge name

Please include country code

Your email address is required for a copy of the conference attendee list

� Please do not include my name for special offers or promotions from CMG partners.

CMG2005 Attendance Options: Please mark the appropriate option and price. Complete this form and return it to the address at the top of the form. To take advantage ofthe Early Bird discount rate, your registration must be received at CMG Headquarters by the date listed below.Registration Fees: Options 1-2 and 6-7 include attendance to all sessions, BOFs, Exhibitor Presentations, Exhibit access Tuesday-Thursday, Breakfasts and LunchesTuesday-Friday, and PARS (Options 3 & 4 include the same functions for those days registered). Options 1, 5 & 6 include attendance to the Sunday Workshops, Breakfast andLunch on Sunday. All registration options (EXCEPT options 6, 7, and 8) include a one year CMG membership for the year immediately following the conference.Refund Policy: All cancellations must be in writing on company letterhead and must be received at CMGHQ by November 18, 2005. CMG will deduct the cost of a one-yearmembership plus a $150 processing fee from all refunds. No refunds will be issued after that date. Cancellation: If the CMG conference cannot be conducted due to acts of God, war, government regulation, disaster, strikes, civil disorder, curtailment of transportation facilitiesor other emergencies making it inadvisable, illegal or impossible to provide the facilities or to hold the meeting, each prepaid registrant will receive a copy of the conferenceProceedings and any other gifts. Membership dues of pre registrants and other fixed conference expenses will be paid from the pre registration funds. Any remaining funds willbe refunded to pre registrants. CMG is not responsible for any other costs incurred by pre registrants in connection with the conference.

Special Meal Needs: � Vegetarian � Kosher

Option 1 Full Week Conference & Sunday Workshops $1,650.00� $1,850.00� $1,900.00� $2,100.00�Option 2 Full Week Conference Only $1,375.00� $1,575.00� $1,625.00� $1,825.00�Option 3 One Day Conf. (exhibits when open) � Tue. � Wed. � Thu. � Fri. $475.00� $525.00� $550.00� $600.00�Option 4 Two Day Conf. (exhibits when open) � Tue. � Wed. � Thu. � Fri. $775.00� $875.00� $900.00� $1,000.00�Option 5 Sunday Workshops Only $475.00� $475.00� $475.00� $475.00�Option 6 Corporate (shared) Badge - Full Week & Workshops N/A $1,850.00� N/A $2,100.00�Option 7 Corporate (shared) Badge - Full Week N/A $1,575.00� N/A $1,825.00�Option 8 Guest Registration: Breakfast and PARS tickets $350.00�

CMG2005 CD-Rom Proceedings: All registered attendees will receive on site with their collectibles FREE ⌧

CMG2005 Printed Proceedings - 2 Volume $60.00�GRAND TOTAL

Registration after 10/14/05Member Non-Member

Early Bird by 10/14/05Member Non-Member

Payment Options: Full payment in US dollars, check, money order or credit card must accompany the registration form or the registration cannot be processed.� Check Enclosed (payable to The Computer Measurement Group, Inc.) � VISA � Mastercard � American Express

Credit Card #:________________________________________________ Expiration Date:______________ Zip Code:__________________

_______________________________________________ ______________________________________________________________Name of Cardholder (please print) Signature of Cardholder (signifies authorization to charge credit card account)Note: Monies paid to the Computer Measurement Group are not deductible as charitable contributions but may be deductible as ordinary & necessary business expenses. See your tax consultant for advice.

37

CM

G2

00

5•

Or

la

nd

o

38

HOTEL & REGISTRATION FORMS

REGISTRATION OPTIONS

OPTION 1 FULL CONFERENCE & SUNDAY

WORKSHOPS

Includes entry to all CMG2005 sessions, BOFS,exhibitor presentations, exhibit hall access,meals/PARS, Proceedings, Sunday Workshops, and2006 CMG membership.

OPTION 2 FULL WEEK CONFERENCE

Includes entry to all CMG2005 sessions, BOFS,exhibitor presentations, exhibit hall access,meals/PARS, Proceedings, and 2006 CMGmembership. (Sunday Workshops are not included)

OPTION 3 ONE-DAY CONFERENCE

Includes entry on the day specified to all CMG2005

sessions, BOFS, exhibitor presentations, exhibit hallaccess, meals/PARS, Proceedings, and 2006 CMGmembership. (Sunday Workshops are not included)

OPTION 4 TWO-DAY CONFERENCE

Includes entry on the days specified to all CMG2005

sessions, BOFS, exhibitor presentations, exhibit hallaccess, meals/PARS, Proceedings, and 2006 CMGmembership. (Sunday Workshops are not included)

OPTION 5 SUNDAY WORKSHOP ONLY

Includes entry to Sunday Workshops includinghandouts for the Sunday Workshops, breakfast andlunch on Sunday, Proceedings, and a 2006 CMGMembership.

OPTION 6 CORPORATE BADGE – FULL

WEEK AND SUNDAY WORKSHOPS

Each Corporate Badge specifies a company name, notan attendee name. Includes entry to all CMG2005

sessions, BOFS, exhibitor presentations, exhibit hallaccess, meals/PARS, Proceedings, and SundayWorkshops. This provides local companies andCMG2005 exhibitors the opportunity to send moreassociates to a variety of conference sessions atdifferent times during conference week. There is no

2006 CMG membership with this option.

OPTION 7 CORPORATE BADGE – FULL WEEK

Each Corporate Badge specifies a company name, notan attendee name. Includes entry to all CMG2005

sessions, BOFS, exhibitor presentations, exhibit hallaccess, meals/PARS, and Proceedings (SundayWorkshops are not included). This provides localcompanies and CMG2005 exhibitors the opportunity tosend more associates to a variety of conferencesessions at different times during conference week.There is no 2006 CMG membership with this option.

OPTION 8 GUEST REGISTRATION

Includes Full Breakfast and evening PARS. Guests areonly permitted in Exhibit Hall on Thursday.

CANCELLATION & REFUND INSTRUCTIONS

All cancellations must be in writing on companyletterhead and be received at CMG HQ by November19, 2005. CMG will deduct the cost of a one-yearmembership and a $150.00 processing fee from allrefunds.

SUBSTITUTION INSTRUCTIONS

Requests must be in writing on company letterhead,identifying the pre-registered attendee and thesubstitute. The substituted individual must complete aregistration form. You can make a substitution atanytime. CMG allows only one substitution perregistration.

3 EASY WAYS TO REGISTER FOR

CMG2005

1. Register online

Visit the CMG homepage, www.cmg.org. Click the “Register for CMG2005” button, complete the forms, and submit. Online registration requires a valid Visa, Mastercard, or American Express account.

2. Register by fax

Complete the CMG2005 Registration Form and fax it back to CMGHQ at:856-401-1708.

3. Register by mail

Complete the CMG2005 Registration Form and mail it back to:Computer Measurement Group, Inc., P.O. Box 8500-5545, Philadelphia, PA 19178-8500.

For more information, call:

1 800 4ForCMG or 856-401-1700

DE

CE

MB

ER

• 4-9

TUESDAY SESSIONS



Learning to Play NicelyEric Pulier, SOA Software, Inc

By their very nature, Service-Oriented Architectures (SOAs) tend toblur traditional boundaries in large organizations. In the currentenvironment, most IT initiatives involve the coordination of suchdistinct groups as software development, network operations, security,architects, as well as line of business managers. Not so anymore.With the SOA potentially creating reusable software code that must beaccessed dynamically by composite applications, both inside andoutside the firewall, the traditional roles and responsibilities of IT havebeen forever changed, or even erased. In this presentation,technology entrepreneur and SOA visionary Eric Pulier will explore theways in which the SOA pushes the boundaries of IT and look at howlarge organizations can teach their disparate IT groups to “play nicely”and evolve into a successful SOA culture.

Chair: Richard S. Ralston, Humana Inc.



Introduction to Data Center Markup

Language (DCML)Chris Molloy, IBM

As labor costs continue to rise, IT environments are installingautomation to convert labor based tasks. There are heterogeneousniche products that have been proven very effective. The Data CenterMarkup Language (DCML) standard has been proposed by aconsortium of companies to allow one to codify the current state of anIT environment, and provide a metalanguage for defining the policiesfor how that environment should be run. This paper discusses DCML,and takes a look at how performance information can be defined in acommon format so that disparate products may use the information.

39


CM

G2

00

5•

Or

la

nd

oWEDNESDAY SESSIONS

SESSION 401 WEDNESDAY 8:00 AM - 9:00 AM

PAPER 5412 No PaperWINDOWS

The China EraAmy D. Wohl, Wohl Associates

China is on the rise. Sometime in this century America will need torecognize it as (at least) an economic Super Power.

But what does China’s evolution from an enormous, largely agrarian,economy to an international investor, an educational powerhouse, anda manufacturing giant have to do with the business of informationprocessing? The answers are at once complex, disturbing, andexciting. It is one part scary (we’ll loose jobs, even very skilled jobs toChinese working remotely from the U.S.); one part challenging (theyare educating their students much better in science and engineeringthan we are); and one part empowering (they are eager investors in theU.S. and excited to partner with U.S. firms on projects in China).

(See complete abstract on page 6)

Chair: Rudy J. Waldner, IBM Corporation



SOA and the Social Order - City

Planning, Post Office & Business

ProtocolsDr. Jeffrey P. Buzen, Independent ConsultantDr. Annie W. Shum, BMC Software

The software industry is abuzz with excitement over promises thatService Oriented Architecture (SOA) will provide unprecedentedbusiness value by containing IT complexity, aligning IT with business,and facilitating IT agility – while also spurring a momentous paradigmshift towards virtualizing and uniformly connecting disparate softwareprograms at the business logic layer. The key to achieving these goalsis the rise of a horizontal service infrastructure layer/businessapplication framework. This nascent but pivotal concept is illustratedthrough familiar metaphors and examples.

Chair: Joan A. Arnold-Roksandich, Hitachi Data Systems


PAPER 5196 TECHNICALMAINFRAME (Z/OS) Z/OS

No Paper

How Do You Do What You Do When

You’re a CPU?Robert R. Rogers, IBM Corp.

If you’ve ever been an assembler programmer, you’ll enjoy this lookinside the modern CPU given by one of its architects. It includes anoverview of the processor cache, the elements of the instructionpipeline and the best way to move data. This version of thepresentation has been updated to describe the superscalar, in-orderinstruction processing of the IBM z990.

Chair: Mel Boksenbaum, Hitachi Data Systems


PAPER 5413 ADVANCEDFUNDAMENTALS/CORE COMPETENCY WINDOWS NT/XP/ETC.

No Paper

The Application Performance Index

(Apdex) Standard+Peter Sevcik, NetForecast, Inc.

Application performance measurement tools have several limitations.Each has its own definition of performance, they produce too manyconfusing or conflicting numbers, and they lack a simple summaryresult that can be understood by corporate management. A consortiumof more than 20 companies is developing the Application PerformanceIndex (Apdex) to bridge the reporting gap by specifying a uniform wayto measure and report on the user experience. This paper describesthe index and provides an update on its standardization.

Chair: Stephen Guendert, McDATA


PAPER 5110 INTRODUCTORYSTORAGE NON-SPECIFIC

Smoke and Mirrors – A Survey of

Remote Replication TechnologiesCharles T. McGavin Jr., EMC CorporationTony Mungal, EMC Corporation

Increasing business and regulatory pressures are forcing more andmore enterprises to consider remote replication (replication of datafrom one storage array to another). Fortunately, there are moreoptions than ever before. This paper will survey the current landscapeof remote replication technologies, examining the technologies andanalyzing their advantages and disadvantages.

Chair: Gregory A. Korbecki, ABN AMRO Services Co.



Reflection: The Next Big Thing After

SOANick Gall

Emerging concepts such as the following are actually differentmanifestations of one emerging trend: Aspect-Oriented Architecture,Dynamic Systems Initiative, Model Driven Architecture, REST(Representational State Transfer), Service-Oriented Architecture, andWeb & Web Services Architecture (including Web Services DistributedManagement). We call this emerging trend “Decentralized ReflectiveArchitecture”. A reflective architecture is one that is Self-describing,Self-discovering, Self-relating, Self-modifying, Self-similar (fractal).

Twenty-plus years of academic research into “reflection” is now readyto commercially harvest. In fact, it is already happening. Imagine thatall information was literally distributed through a global informationspace with possibly millions of replicas kept in varying degrees of syncusing gossip and epidemic protocols. Such a global information spaceis becoming an absolute requirement for EPCGlobal, Network CentricWarfare, global sensor networks, etc. This presentation will:

• Explain the concepts of decentralized reflective architecture are emerging from SOA and WS-* principles

• Show how it can be applied to solve business problems • Provide examples of vendor architectures that embody it • Discuss several proof of concepts, possibly including a brief demo

40

DE

CE

MB

ER

• 4-9

WEDNESDAY SESSIONS

41



Workload Generation: Does One

Approach Fit All?Alexander Podelko, Hyperion Solutions

A must task in load testing is workload generation: how to apply a loadto your system. It is important to understand all possible options; asingle approach may not work in all situations. The main choices areto generate workload manually, to use a load testing tool or to create aprogram to generate a load. Many tools allow you to use different waysof recording/playback and programming. This paper discusses prosand cons of each approach mainly based on experience withdistributed business applications.

Chair: Mattie McKnight, SAS Institute


PAPER 5069 HOW-TOFUNDAMENTALS/CORE COMPETENCY Z/OS

Reactive Capacity Planning – An

AlternativeDick Arnold, JP Morgan Chase Bank

Since the practice of capacity planning first spun off from theperformance measurement and tuning area around 1980, it hasgenerally evolved into fairly commonly used techniques. However, anumber of things have changed over the last 20 years . There is lessstaff to do the very labor intensive work, CPU upgrades are lesscomplex and upgrade costs are much lower. But, we are still doing thejob the same way. There is another way that is faster, takes less effortand costs less. And, even though there are some tradeoffs and itrequires some management buy in, it may work for you.

Chair: Thomas E. Bell PhD, Rivendel Consultants Inc

SESSION 421 WEDNESDAY 10:30 AM - 12:00 PM

PAPER 5178 INTRODUCTORYUNIX/LINUX UNIX/LINUX

No Paper

Linux Performance TuningJaqui Lynch, Mainline Information Systems

This presentation only session will go into the basics of performancetuning on a Redhat Linux system. Tuneables, as well as somearchitectural issues, will be discussed.

Chair: Sean Shepley, UKCMG



No Paper

WebSphere Application Server for

z/OS Version 6 - Performance TuningGlenn R. Anderson, IBM

Now that you have installed WebSphere Application Server for z/OSVer 6, how do you know that your system has been tuned for optimalperformance? This session will provide performance tuningrecommendations for WebSphere on z/OS, including configurationconsiderations, WLM settings, WebSphere tuning, JVM tuning, andz/OS system tuning . The session will also provide information abouttools and techniques which can be used for performance analysis,such as RMF and various tracing options. We will also look at howWebSphere V6 on z/OS interacts with Enterprise Workload Manager.

Chair: James Yaple, US Dept. of Veteran Affairs


PAPER 5193 TECHNICALWINDOWS WINDOWS NT/XP/ETC.

Virtual Memory Constraints in 32-bit

Windows: an UpdateMark B. Friedman, Demand Technology Software

This paper discusses the signs that indicate a machine is sufferingfrom a virtual memory constraint in 32-bit Windows. It also discussesoptions to keep this from happening, including (1) changing the way32-bit virtual address spaces are partitioned into private and sharedranges, (2) settings that govern the size of system memory pools, (3)hardware that supports 37-bit addressing, and (4) hardware thatsupports 64-bit addressing but can still run 32-bit applications incompatibility mode.

Chair: Michael B. Marcus, Affiliated Computer Services

CM

G2

00

5•

Or

la

nd

oWEDNESDAY SESSIONS


PAPER 5136 TECHNICALSTORAGE WINDOWS NT/XP/ETC.

A Management Framework For

Petabyte-Scale Disk StorageMichael A. Salsburg, UnisysDavid Lifka Ruth S. Mitchell, Cornell Theory Center

An explosion of data, along with new requirements for corporategovernance, are generating requirements to build petabyte-sized disksubsystems. Given the current costs and complexities inadministration for terabyte-sized systems, new innovations inmanagement and monitoring are needed to scale beyond a petabyte.We are currently constructing a disk subsystem that will exceed apetabyte within the next few years. This paper presents, from apractitioner’s viewpoint, the design and implementation of amanagement framework that is architected to scale beyond a petabyteof storage.


Database Disk to Disk Backups Using

ATA DiskKathleen N. Hodge, Storage Technology Corporation

As the cost of disk continues to decrease, disk-based backups havebecome more attractive. Although tape backups still have their placein data archival and off-site vaulting, they are not a very efficientmethod of data recovery. Database backups are staged on disk inpreparation for a tape backup routine to copy to tape. The stagedbackup can be stored on inexpensive ATA disk relieving expensiveprimary disk storage. This white paper describes ATA disk capacitiesavailable for backup solutions, and also presents storage strategies forbackups throughout the lifecycle of a database.

Chair: Edna B. Morrison, RBC Financial Group


PAPER 5111 TECHNICALNETWORK NON-SPECIFIC

Intrusion Detection/Prevention

Devices — Are They Protecting Your

Network - or Hampering It?Neil Carter, Spirent Communications

IDS/IPS devices are burdened with serious and increasing challenges.Performance of these devices is affected by factors such asconfiguration, traffic types, and security they provide. Companieslooking to deploy IDS/IPS should measure the performance, reliabilityand overall security. How do different protocols/applications (normaltraffic) used in networks combined with virus, DDoS and spam attacksaffect IDS/IPS performance? This paper will delve into the challengesof IDS/IPS deployment, using real life situations and show the impactof performance testing and measurement.


Want to Know WHY Response Time

is So Long? Listen to the Wire.Jack B. Woolley, Kaiser Permanente

Kaiser Permanente has been using application network traffic analysisas part of our application performance problem determination.Unencrypted TCP/IP network activity often identifies the portion(s) ofthe application that result in poor overall performance.

Training in this technique has the ability to identify poor performingapplication modules/objects, CPU constrained components,implementation configuration issues, and specific poor performingapplication SQL.

This paper presents the basic techniques of this type of problemdetermination, along with several real life examples.


PAPER 5179 ADVANCEDFUNDAMENTALS/CORE COMPETENCY NON-SPECIFIC

Is It Time for Capacity Planners to

Hang Up Their Cleats?G Jay Lipovich, BMC Software, Inc

Historically, capacity planning has delivered measurable value to theorganization in the form of reduced costs, deferred acquisitions, andservice level delivery, deliver a return on investment (ROI) for theorganization. However, fundamental changes in the underpinnings ofcapacity planning ROI threaten the traditional value of capacityplanning, and raise the question of whether it is still worth it to engagein capacity planning efforts. This paper examines these changes andtheir impacts and then considers actions capacity planners might taketo adapt to the new reality.

42

DE

CE

MB

ER

• 4-9

WEDNESDAY SESSIONS

43


PAPER 5031 INTRODUCTORYFUNDAMENTALS/CORE COMPETENCY NON-SPECIFIC

Using Six Sigma to Define the Focus

of Software Performance EngineeringKevin Mobley, Fidelity Information Services, a division ofFidelity National F

Software Performance Engineering Failure Modes and Effects Analysis(SFMEA) is a method to assess risk and determine the top businessprocesses (bps) to focus on during software architecture and design,performance simulation and optimization, and performancemanagement. Specifically SFMEA identifies how bps violate softwareperformance anti-patterns and combines this risk with the bps’frequency, current control plan and the voice of the customer. Theproduct of these dimensions is the risk priority number (RPN). The bpswith the highest RPN values defines the SPE focus

Chair: Stephen J. Marksamer, Aetna

SESSION 431 WEDNESDAY 2:00 PM - 3:00 PM

PAPER 5403 NON-SPECIFICFUNDAMENTALS/CORE COMPETENCY No Paper

Workload Characterization for Parallel

Processing EnvironmentsDr. Jeffrey P. Buzen, Independent ConsultantBoris Zbitsker, BEZ Systems

Workload characterization represents a fundamentally important stepin the process of understanding, managing and modeling theperformance of any computer system. After reviewing some universalworkload characterization issues that are common to all systems, thispaper discusses the specific issues, measurement sources and “whatif” questions associated with three distinct parallel processing database environments: Oracle 10g, DB2 UDB ESE and Teradata V2R5.Architectural differences are shown to have an important impact on theperformance of applications running in these three environments.

Chair: Michael S. Recant, MGS, Inc.



zSeries Capacity Management - a

True StoryTony Ruberry, John Lewis Partnership

In September 2003, five months after our previous mainframe upgrade,John Lewis added another two processors to our zSeries machine.Within days the peak hour CPU reached 94%; and we hadn’t evenreached our busiest time of year! Fifteen months later, we are runningwith the same capacity, and have provided excellent service in the runup to Christmas 2004, despite an increase of over 25% in businesstransactions. This paper tells the tale of how, by a mixture of good luckand good judgement, John Lewis have moved from reactive panic tocontrolled and effective capacity management.

Chair: Brian M. Johnson, Bank of America


PAPER 5414 WINDOWS NT/XP/ETC.WINDOWS No Paper

Performance Baselining,

Benchmarking and Monitoring for

Microsoft SQL ServerKevin Kline, Quest Software

It’s impossible to truly tune SQL Server without a good understandingof its performance profile and workload. This session describestechniques and procedures to assess the baseline performance profileof a SQL Server, benchmark the performance of the server undervarying work loads, and how to continuously monitor the server formaximum availability and performance. This session will help you:

1. Learn how to set up baselines, benchmarks, and monitoring2. Understand the benefits and drawbacks of the two main

approaches to monitoring3. Learn major PerfMon counters that are important for baselining,

benchmarking, and monitoring 4. Discuss useful methods for ongoing 24x7 monitoring on SQL

Server boxes.

Basic database administration skills and Windows Server OSmanagement skills are prerequisites.

Chair: Barry Stanton, Netco Government Services Inc.


PAPER 5102 TECHNICALSTORAGE UNIX/LINUX

I/O Performance Characteristics for

Volume Managers on Linux 2.6

ServersDan W. Yee, VERITAS software CorporationXianneng Shen, Veritas SoftwareRandy Taylor, Veritas Software

This study investigates the I/O performance characteristics of rawvolumes in a Linux 2.6 environment running on 64-bit processors (AMDOpteron and Intel Itanium 2). Both sequential and random I/O aretested on RAID-0, and RAID-1+0 devices. The volume managers usedare VxVM, Linux LVM2, and Linux MD. Synthetic workloads are usedin the study. Measurement results are reported.

Chair: Xianneng Shen, Veritas Software

CM

G2

00

5•

Or

la

nd

oWEDNESDAY SESSIONS

44



Radio-Frequency Identification at

Enterprise ScaleKen Traub, ConnecTerra

Radio-frequency Identification (RFID) is being rapidly adopted withinmany industry verticals, driven in part by mandates from leadingretailers, regulatory agencies, and the federal government. Adoptionwill have significant impact on IT infrastructure for enterprises. Forexample, a nationwide retail chain may require ten million RFIDreaders, representing a manageability problem and a source of newdata, orders of magnitude larger than anything deployed today. Thistalk will present the fundamentals of large-scale RFID deployments,with particular attention on the enterprise IT challenges. Topicsdiscussed include architecture, software standards, management, andscalability.

Chair: Annette S. Kakazu, Acxiom


PAPER 5424

CMG Australia Best Paper:TBD

TBD



Wholesale Distributed Capacity

PlanningIrving Smith, Bank of AmericaRussell W. Burns, Bank of America

In today’s multi-tiered mass server infrastructure, it is highly desirableto be able to provide capacity reports in a timely, proactive and costeffective manner. This paper presents a methodology for producing alevel one automated capacity planning report for a largeheterogeneous, distributed systems server population. The paperdescribes the forecasting techniques employed to generate the reportand the capacity planning process that it supports.



QNS - an Online System for the Study

of Queuing ModelsDr. Jozo J. Dujmovic, San Francisco State UniversityHemamalini Sankar

QNS is an online system for the self study of the most importantqueuing models. QNS is publicly available on the Internet. It offers acomplete educational support including: (1) theoretical presentation ofmaterial, (2) queuing theory models, (3) graphical simulator withqueuing network animation, (4) laboratory experiments based onnumerical solver, (5) quiz subsystem with automatic grading, (6)control system with GUI, and (7) remote access support. The paperpresents the design and implementation of QNS using JSP, HTML,Java, and Tomcat.

Chair: Martin D. Brake, IBM Global Services


PAPER 5052 HOW-TOMAINFRAME (Z/OS) Z/OS

Migrating to z-990 - A User

ExperienceIan C. Baldwin, Barclaycard

The author has recently been through the task of migrating hisapplications and services from z-900 to z-990 technology. This papergives the authors experiences, processes, suggestions and lessonslearnt to aid like-minded IT professionals with similar migrations and toaid in truly understanding what the z-990 can achieve.

Chair: Xianneng Shen, Veritas Software


PAPER 5421 WINDOWS NT/XP/ETC.WINDOWS No Paper

Panel: IT’s Role in Business

Performance ManagementMichael A. Salsburg, PhD, Unisys Corporation

Business Performance Management raises performance analysis andmanagement to a new level of the enterprise. It is concerned withways to improve the overall business, using Key PerformanceIndicators (KPIs) to monitor enterprise-wide performance as viewed bythe executive decision makers. When necessary, strategy andbusiness processes need to be adjusted to improve performance andachieve the goals of these indicators. How will we apply our expertiseand discipline to have a positive impact on these KPIs? Whatemerging IT technologies can be brought to bear on this higher level ofmanagement? The panel of specialists will explore these and otherquestions regarding the impact of IT on Business PerformanceManagement.

Chair: Michael Gardner, RBC Financial Group


PAPER 5115 ADVANCEDSTORAGE NON-SPECIFIC

Information Classification and

Service Level Objectives for

Information Lifecycle ManagementRobert Rogers, Application Matrix

Information Lifecycle Management (ILM) is the alignment of storageresources services to business goals and objectives. It is criticallyimportant to have an ILM methodology that associates businesselements with their resources. The alignment of resources depends onunderstanding what is important, what the business needs are, andwhat corporate and regulatory requirements affect the business (e.g.,Sarbanes-Oxley, and HIPAA). This paper focuses on those techniquesand strategies for classifying information and data.

DE

CE

MB

ER

• 4-9

WEDNESDAY SESSIONS

45


PAPER 5181 ADVANCEDNETWORK UNIX/LINUX

Performance Tuning of Gigabit

Network InterfacesProf. James Westall, Clemson UniversityRobert Geist, Clemson UniversityJames J. Martin, Clemson University

As Ethernets have attained gigabit speeds, it has become obvious thatNIC and device driver designs that work well at slower speedsintroduce performance problems at gigabit rates. In this paper wepresent a systematic evaluation of the effects of frame size andinterrupt coalescing strategy on the performance of three benchmarkworkloads on a gigabit LAN. We show that the use of large frames andinterrupt coalescing can produce significant performance benefits, butthat the proper degree of interrupt coalescing is strongly dependentupon the characteristics of the workload.

Chair: Catherine S. Nolan, Bank of America



Through the Prism of Fractals: Why

SOA Should Reflect the Natural OrderDr. Annie W. Shum

Service Orientation is emerging as the fourth wave of the computingparadigm shift because it promises to enable broad-scaleinteroperability and unprecedented business agility in a service value-net (ecosystem). Containing IT complexity and aligning IT withbusiness through a set of sound and robust design principles arepivotal to the transformational power of Service Oriented Architecture(SOA). This paper looks for insights into containing IT complexity bystudying the time-tested tenets and dynamics of complex fractal-likeforms that abound in Nature.



Capacity Planning for Shared

Middleware Environments; A

MethodologyChandra Lanka, Bank of America

Sharing middleware application environments has gained significancein the recent past. Heterogeneous mix of transactions that run in theshared environments increases the complexity of both performancemanagement and capacity planning. Some transactions consumemore CPU power and memory than others. Different transactions peakat different times of the day. If and when all transactions peak at thesame time, there may be a ‘’perfect storm’’. In this paper, I discuss theissues involved in capacity planning for a shared webMethodsenvironment and present a methodology for capacity planning.

Chair: Bryan Drake, VITA

CM

G2

00

5•

Or

la

nd

o

46

SESSION 501 THURSDAY 8:00 AM - 9:00 AM


No Paper

z990 Performance and Capacity

Planning IssuesCheryl Watson, Watson & Walker, Inc.

Several installations believe their new z990 processors are notperforming as expected. Are their beginning expectations unrealistic?Do the new configurations require a different level of tuning? Are themachines not performing according to LSPR expectations? You mayfind that one or more of these is true in your case. Whether you havez990 processors currently installed, or are planning on ordering them,this is an extremely important session. The session is given by CherylWatson, who has an intense interest in these new machines andextensive experience in comparative performance studies.



Taking FICON to the Next Level-

Cascaded High Performance FICONStephen R. Guendert, McDATA Corporation

FICON technology has been in place for five years now, and in the pasttwo years significant advances have been made. One of thoseadvances is cascaded FICON. This paper will discuss FICONcascading in depth: what it is, what its benefits are, what the importantz/OS considerations are when implementing, how to implement, andfinally what are the performance considerations and how to get betterperformance via optimizing buffer to buffer credits. The author will thengo through a real life example of modeling and testing that was doneto optimize a cascaded FICON environment.


PAPER 5135 TECHNICALWINDOWS WINDOWS NT/XP/ETC.

Understanding and Interpreting SQL

Server Performance CountersJeffry A. Schwartz, Unisys Corp.

SQL Server makes many performance counters available. Howevernumerous explanations provided via the Windows PerformanceMonitor simply restate the name of the counter or provide crypticexplanations. This paper discusses the interpretation of severalperformance counters that have proven useful in performance studies,elaborates on their Windows Performance Monitor explanations, andproposes potential courses of action.

Chair: Sidney W. Soberman, H W Wilson



Encryption Primer: An Introduction to

Data ProtectionCathy Nolan, Bank of AmericaJohn C. Besci Prentice O. Dees

From an enterprise-level perspective, encryption is the most-oftenmentioned option for protecting stored confidential or proprietaryinformation today. This paper will give you an overview of encryption,including what problems encryption can solve and what problems itmay introduce. Also discussed will be the challenges of developing anencryption strategy for protecting data-at-rest in enterprise-levelenvironments, as well as performance impacts of encrypting data andkey management issues.


PAPER 5408 No PaperNETWORK

Measuring End User Response Time -

With a Passive Network ProbeBernie Davidovics, SeaNet Technologies

One of the most vital yet elusive measurements has always been the“End User Response Time”. This presentation discusses recentbreakthroughs in using passive network probes to effectively measurethe end user response time of not only the low level components ofresponse time - but also the logical business transactions. Thispresentation describes how traffic collected on the network can bedecoded and analyzed to provide the most vital end usermeasurements - without the need to install server side or client sideagents.



Software Performance Engineering

Considerations in Unreliable

Computing EnvironmentsDr. Pierre M. Fiorini, University of Southern MaineYiping Ding, BMC Software

In this paper, we discuss software design issues that should beconsidered whenever jobs execute in unreliable computingenvironments. Specifically, we show that if proper checkpointingmechanisms are not properly implemented, then under certainconditions completion times of applications executing on the systemexhibit properties of heavy-tail or power-tail distributions, which canlead to unpredictable and long completion times.

Chair: Jerry L. Rosenberg, SRM Associates Ltd

THURSDAY SESSIONS

DE

CE

MB

ER

• 4-9

THURSDAY SESSIONS

47



Survivor - The Corporate JungleDenise P. Kalm, BMC Software

The IT world has changed. With job losses exceeding 400,000 in thelast four years, survival in this New Age requires much more than justputting in a good day’s work. The tools and strategies you needinclude: career assessment, ‘managing up,’ personal public relations,networking and much more. Learn how to outwit, outplay and outlastthe competition – to thrive, instead of just survive.

Chair: Barry N. Sokolik, Charles Schwab



DB2 CPU and Response MetricsNed A. Diehl, ISM

DB2 requests originate from a variety of diverse sources includingbatch, CICS, DDF and SAP. Related DB2 CPU usage can be recordedin RMF, DB2, SMF 30, and other subsystem records. Proper selectionand interpretation of these values will vary with transaction source,DB2 environment, product levels, and analysis objectives. Analystsmust be careful to include all desired values and avoid multiplecounting of the same logical utilization. This paper will discuss thesources and analysis of DB2 CPU metrics and correspondingresponse times. Examples will include CICS, DDF, and SAP.


PAPER 5164 INTRODUCTORYFUNDAMENTALS/CORE COMPETENCY WINDOWS NT/XP/ETC.

Measuring Performance in the Lab

and Validating it in ProductionEllen M. Friedman, SRM Associates, Ltd.

You have a script to run which represents your productionenvironment. What do you measure, what do you look at to make sureit is representative, and examining performance? What to look forwhen running the tests, how to compare the results between testiterations and re-evaluating performance after hardware/softwarechanges. When you are done, how do your tests compare in your firstassessment in a production environment? The case study presentedwill be SQL server under Windows


PAPER 5128 TECHNICALHOT TOPICS NON-SPECIFIC

Redefining Capacity Planning for

Grid ComputingMarina Cismas, Bank of America

Grid computing is an emerging concept that will change the way weuse and think about technology. This paper will explore the questionssuch as how the role of capacity planning will change in respect to therise of grid technology and why the field is becoming ever more critical,despite rapid hardware price-performance improvements. The paperwill also discuss the benefits and challenges that capacity planninganalysts will face when working with grids and explore the newcapacity planning objectives and innovative performance improvementstrategies promised by the grid technology.



Closing the Gaps – Understanding

Capacity SummarizationLinwood Merritt, Bank of America

One of the first steps in the capacity planning process is thesummarization of capacity/performance data. This activity aggregatesdetailed event-level data into a manageable number of workloads andintervals. Interval data can be summarized into average, peak orpercentile figures. This paper discusses summarization techniquesand intervals, and the gaps between results from different choices ofeach.

Chair: Ivan L. Gelb, Gelb Information Systems Corp



Introduction to Simple Network

Management Protocol (SNMP)

Version 3Laura Knapp, IBM

Simple Network Management Protocol has emerged as the de factostandard for the management of networks. Growing from its initialfocus of managing devices in a TCP/IP network, this protocol is nowused by networking elements, operating systems, and applications,allowing a common management infrastructure to be put in place.

One of the gravest exposures in today’s business networks comesfrom the SNMP management software. The lax security makes this ahacker’s dream come true and allows entry points to private businessnetworks. But fear not, SNMP v3 has the functions designed to secureyour management flows. This is even more important as thedemarcation lines blur between your business and the Internet.

This session explores SNMP, providing you the basics and thearchitecture enhancements in SNMP v3 to support new requirementsand explores how you can use it to manage your overall end-endsystem.



A Performance Model Web ServiceDr. Connie U. Smith, Performance Engineering ServicesCatalina M. Lladó, Universitat Illes BalearsRamon Puigjaner, Universitat de les Illes Balears

Performance Engineering uses multiple performance assessmenttools depending on the state of the software and the amount ofperformance data available. This paper demonstrates how WebServices can be used to facilitate the use of modeling tools in a plug-and-play manner thus enabling the use of the tool best suited to theanalysis. The paper describes the design and implementation of aprototype Web Service for a performance modeling tool. Additionally, itshows experimental results that prove the viability of such a WebService.

CM

G2

00

5•

Or

la

nd

oTHURSDAY SESSIONS



Benchmarking 101James A. Yaple, Austin Automation Center Jennifer Sutherland, State of Wisconsin - DHFS

This paper will explore benchmarking using tools and productsavailable in the commercial marketplace. The material will includeoverviews of commonly used open systems benchmarks, whatbenchmarks are designed to measure, how vendors use and misusebenchmarks, how customer organizations can use benchmark resultsand the value of customers establishing an internal benchmarkingprocess. This paper presents published results from several currentindustry-standard benchmarks and analyzes them.

SESSION 521 THURSDAY 10:30 AM - 12:00 PM


No Paper

The Mechanics of Developing a High

Quality Capacity PlanKathy Walsh, IBM

There are many steps to developing a successful capacity plan. Errorsin any of these steps can cause the resulting capacity plan toincorrectly set an expectation for a new processor or LPARconfiguration change. This session will review the actual steps in theplanning process which often cause these problems with suggestionson how to avoid potential pitfalls through the use of new tools.



Unveiling of DB2’s DDF: SQL

Revealed via the Gestalt PerspectiveThomas A. Halinski, Compuware Corporation

DB2’s Distributed Data Facility’s (DDF) veil has been lifted from theeyes of the eBusiness applications personnel. With the ability to viewand measure the performance of DDF SQL, which runs outside of thenormal DB2 address space, IT personnel finally have insight into theirsystems. This paper will discuss: the evolution of using DDF - from aGestalt perspective; DDF’s components and how they interact withMVS - key elements in understanding this process. Using special toolsto enter the former “black box” of DDF’s address space, its SQL isrevealed and thus tunable.



Designing a Fairer Round Robin

Scheduling AlgorithmRichard K. Roehl, Bank Of America

Operating Systems that implement a simple round robin scheduler areinherently inefficient. In this environment tasks that have consumed orconsume more CPU than their peers get equal scheduling priority onthe CPU. By implementing a priority queue on the scheduler you caneasily improve performance of processes that require short bursts ofCPU while continuing to service the processes with higher usagedemands.

The example used is based on the Minix operating system. However,this scheduling algorithm can be applied for any operating system thatimplements a round robin scheduler.


Capacity Planning for UNIX System

MetricsSerge Tessier, Volvo

In recent years, the Volvo AB group realized many acquisitions thatimpacted strongly the overall IT processes. This triggered the need forglobal troubleshooting tools and procedures, as well as a global web-based monitoring facility allowing worldwide servers’ diagnosticallyproactive survey. Starting our project, we were seeking for a solutionaiming at presenting a consolidated vision of our UNIX servers througha single portal for system administrators, managers and, in some way,customers. The scope was to get resource utilization for a UNIX servervia home-made development.


PAPER 5077 TECHNICALSTORAGE Z/OS

Proper Sizing and Modeling of

ESCON to FICON MigrationsStephen R. Guendert, McDATA Corporation

This paper will discuss the aspects and best practices of sizing anESCON to FICON migration to result in a high performance FICONarchitecture. We will discuss disk, tape and CTC considerations. Thepaper will then demonstrate these principles applied to real worldexamples of FICON migrations using data from actual client sizingsusing modeling tools.

48

DE

CE

MB

ER

• 4-9

THURSDAY SESSIONS


PAPER 5157 INTRODUCTORYNETWORK Z/OS

One-Minute TCP Stack AnalysisNalini J. Elkins, Inside Products

The One Minute Manager suggests three principles for effectivemanagement: one-minute goals, one-minute praisings and one-minutereprimands. These principles apply equally well to TCP networkanalysis. Regularly finding and fixing the problems in one minute ofTCP stack activity can keep your network running smoothly. If you findand fix the problems in the packets transferred in one minute from yourTCP stack every week for a year, you will have found and fixedhundreds of problems.


Underground SPE: Moving from

Performance QA to SPEDr. Mary R. Hesselgrave, How Many? How Fast?

I spent 18 months with a client who asked for a performance test team.By the end of the engagement, we were able to put in place a trueSoftware Performance Engineering (SPE) process that began at therequirements stage, and that included tasks involving systemsengineering, development, support for development tuning, qualityassurance, production monitoring, support for evaluation of proposedproduction configuration changes, and capacity planning.

How was this achieved? This talk discusses the steps we took to makeSPE happen for an organization that did not know it wanted SPE.



Choosing a Load Testing StrategyHeidi A. Gilmore, Segue SoftwareKlaus Fellner, Segue Software

This paper introduces load testing, an important component inoptimizing software quality. Proper load testing can help to mitigate thecosts of poor quality. The paper will explain the importance of loadtesting, when in the software development process and how optimalperformance can be achieved by conducting proper load testing andcapacity planning. In addition to discussing the various strategies forimplementing load testing within an organization, we will also explorethe very real benefits load testing returns to the organization.



Capturing System Data Using Native

CommandsRobert F. Patterson, Bank of America

Our company was paying rather large licensing fees for third-partyproducts for gathering systems data for capacity planning andperformance management. Our focus was to save cost and at thesame time obtain data for mid-range systems like UNIX, Linux, andWindows servers. We wrote scripts to gather our own data usingnative commands.

This paper does comparisons of the various native commands andmetrics/data in AIX, Solaris, HP-UX, Linux, and Windows and presentsan overview of the capture methodology.


Capturing Workload Pathology by

Statistical Exception Detection

SystemIgor A. Trubin, Capital One

The paper describes one site’s experience of using MultivariateAdaptive Statistical Filtering (MASF) to automatically recognize somecommon computer system defects such as run-away processes onone or multiple CPUs and memory leaks. A home-made SEDS(Statistical Exception Detection System) that captures any global andapplication level statistical exceptions was modified to recognize,separately report and alarm about those particular defects.

Chair: Michael L. Knych, International Paper

SESSION 531 THURSDAY 2:00 PM - 3:00 PM


WLM Sysplex ManagementPeter Enrico, Enterprise Perf Strategies, Inc.

What Sysplex considerations does WLM take into account whenmanaging work towards goals? How does WLM manage workloads ina Sysplex, and what interest in the Sysplex does WLM have? Recentlythere have been many questions that revolve around the topic z/OSWorkload Manager (WLM) Sysplex management. The basic conceptsof WLM Sysplex management have not changed much since theintroduction of WLM, but many questions stem from the ever morecommon practice of combining more and more unlike system imagesinto a single Sysplex. This paper explores this important topic.

49

CM

G2

00

5•

Or

la

nd

o

50


PAPER 5209 PANELUNIX/LINUX NON-SPECIFIC

No Paper

Facilitated Discussion: Future of the

Performance Field 2005Rick Lebsack, IBM

Decreases in the cost of hardware is no longer being cited as thereason the performance field will disappear. However will it continue toevolve, or will it wither as other topics capture management attention?The topic of this facilitated discussion will be perceptions of the future,and how individuals should focus their attention to survive and prosper.Opinions of CMG attendees will be interchanged on a number totopics, and senior members of the profession from a academia,business and consultancy will offer their ideas.

Panelists include: Daniel Menasce of George Mason University, FrankBereznay of Kaiser Permanente, Tom Bell of Rivendel Consultants Inc.


PAPER 5122 HOW-TOUNIX/LINUX UNIX/LINUX

Visualization of Performance DataJames Holtman, Convergys

Measuring performance of a system creates a large amount of datathat has to be analyzed, summarized and then used to make decisionson. Many systems look at just textual/reports that summarize by timeof day or business events. It is sometimes hard to understand what thedata is trying to tell you.

This paper will describe a number of ways of visualizing the data interms of charts, graphs and other descriptive techniques. The paperwill show how this can highlight patterns in the data and provide foralternative drill downs on the data.


PAPER 5140 ADVANCEDSTORAGE UNIX/LINUX

Enhancing Web Server Performance

Through the Use of a Drop-In,

Statically Optimal Disk SchedulerProf. Robert Geist, Clemson UniversityJay Steele, Clemson UniversityJames Westall, Clemson University

With version 2.6 of the Linux operating system, disk schedulers havebecome totally modular. This paper provides a performancecomparison of the four native schedulers included with the 2.6 kernel.Also of the authors’ new algorithm (the table-building bus driver) whoseimplementation is, like the others, available in a single C source filethat may be dropped into the kernel. The algorithms are compared ona real system under a workload designed to emulate heavy web servertraffic. The authors’ algorithm is seen to deliver substantialperformance improvements over the other four schedulers.



Measuring Actual and Perceived User

Experience on the WebChris Loosely, Keynote

What do you know about your customers’ experience? How satisfiedare they after each visit to your site? How well does your site performwhen users attempt critical tasks? What frustrations do they have?How likely are customers to make a purchase and return to your site?What impact does your site have on a customer’s impressions of yourbrand? Companies doing business on the Web know that siteperformance is a continually rising target. Customers’ expectations arebeing raised by their experiences when browsing other Web sites andwhen using other Web-enabled (“eBusiness”) applications. In the earlydays, availability was everything; the challenge was to keep your siteup and running. Next came “the 8-second rule” of Web page responsetime, and people began focusing on the need to build and maintain aconsistently responsive Web presence.

At the same time, the measures of site effectiveness have evolved.The simple quantitative metrics available from log file analysis are nolonger sufficient for tracking the user experience on your site andmeasuring its ROI. You need a full-service evaluation program thatestablishes key metrics to assess the effectiveness of your site overtime - as seen by actual users of your site. This presentation willdiscuss how companies are now moving beyond tracking isolatedmetrics like site availability, page response time, traffic volumes, andpage clicks to management approaches that encompass bothquantitative site measurements of site effectiveness, and behavioraland attitudinal data collected from actual uses. The marriage ofquantitative and qualitative information provides a more completeunderstanding of the user experience, and how to improve it.



The Model-Driven (R)evolutionRichard Soley, The Object Management Group, Inc.

All sorts of promises of a revolution in software developmentaccompany the phrase “model-driven” these days. Model DrivenArchitecture, Model Driven Development, Model Driven Enterprise —there must be something to these ideas, but is “model driven” the keyto a revolution, or just the newest buzz word? Will we have tocompletely change the way we develop systems? Is code dead?



8 Great Myths of Software Asset

ManagementMichael P. Swanson, ISAM

Best-in-class companies save tremendously by reducing their softwarecosts via applying the principals of SAM, Software Asset Management.Often companies think that pursuing one course of action vs anotherwill provide the “promised land” of best-in-class cost savings; however,in the author’s research and experience, there are myths associatedwith various strategies. You will see these myths exposed, learn whatto do about them, and how to pursue an enlightened path to lowersoftware costs.

THURSDAY SESSIONS

DE

CE

MB

ER

• 4-9

THURSDAY SESSIONS


PAPER 5501 PANELMAINFRAME (Z/OS) Z/OS

No Paper

Panel: zSeries Performance Q & AIvan L. Gelb, Gelb Information Systems Corp.

If you have a zSeries performance question, this is the panel to ask.Some of the many performance related questions the panel of expertscan answer include: zSeries processors, processor configurations,general Sysplex, z/OS system performance, WLM anything, variableWorkload License Charges, WebSphere, etc... Come prepared withquestions, email them as soon as you can to [email protected], ordrop a written question into the Q&A box you will find at various z/OStrack sessions, or hand your written questions to any z/OS sessionmonitor.

Panelists include: Ivan Gelb of Gelb Information Systems Corp., PeterEnrico of EPS Strategies, Kathy Walsh of IBM Corp., Cheryl Watson ofWatson Walker.


PAPER 5423

CMG Italia Best Paper:TBD

TBD



No Paper

CMG: The Early YearsDavid R. Morley Ian Roome Bill Miller

A lighthearted but very interesting look at the beginnings and rapidgrowth of CMG. How did it begin, why did it grow so quickly, and whatwere some of the more interesting bumps in the road?



Analytic Way for Performance

Management of SANChao Li, Tsinghua UniversityLi-zhu Zhou, Tsinghua UniversityChun-xiao Xing, Tsinghua University

The infrastructure of SAN is complex, and finding a proper method toanalyze and manage the performance of it in an efficient andeconomical way is very important for SAN performance management.This paper proposes an analytic method for performance managementof SAN system, the way to apply the method, and illustrates the methodby analyzing a living example. The comparison of analysis andexperiment results shows that they are in agreement. So we canconsider this method a sound solution for the performancemanagement or control of SAN.


PAPER 5033 ADVANCEDNETWORK NON-SPECIFIC

Workload modeling of Stateful

Protocols Using HMMsDr. Swami Ramany, Network ApplianceRichard Honicky Darren Sawyer, Network Appliance, Inc.

Supporting stateful protocols like CIFS and NFSv4 in a workload model(or a benchmark) requires accurate capture of the order of operationsobserved in various traces as this has a significant impact on theperformance of the storage device.

One common tool for analyzing streams of discrete values that dependon an underlying stateful process is a Hidden Markov Model (HMM).This paper clearly illustrates how HMMs can be used effectively tocapture the state behavior of CIFS and NFSv4 traffic.


PAPER 5060 ADVANCEDHOT TOPICS WINDOWS NT/XP/ETC.

Modeling VMware ESX PerformanceWilliam L. Shelden, Jr., Ph.D., ISM, Inc.

The performance implications of consolidating Windows systems underVMware ESX Server are analyzed using a simulation model. FirstVMware ESX Server overhead is measured based on data fromWindows systems running as virtual machines under VMware ESXServer. Then a simulation model is used to estimate the contention forthe physical CPUs by the virtual CPUs being dispatched from theWindows virtual machines. The simulation results are compared tosome benchmark runs to validate the assumptions of the simulationmodel.



Using Principal Component Method

for Performance Data Compression

and AnalysisDr. Anatoliy Rikun, BMC Software

Modern performance monitoring tools allow the collecting of a vastamount of metrics for many dozens or even hundreds of nodes. In mostcases, there is neither the necessity nor real opportunity to take intoaccount all of the redundant measurements. In this paper wedemonstrate how the method of principal components helps in theanalysis of such multi-dimensional systems, and in particular, how itfacilitates managing the collected data in a compressed form, withoutlosing important information about the system.

51

CM

G2

00

5•

Or

la

nd

oSESSION 603 FRIDAY 8:00 AM - 9:00 AM


No Paper

WLM Caused Pain and PleasureIvan L. Gelb, Gelb Information Systems Corp.

Workload Manager (WLM) can be the cause of intentionally orunintentionally delivered pain, manifested as performancedegradation, or pleasure, manifested as better than expectedperformance. Attendees will learn how to insure that pain and pleasureis delivered as intended. This session will delve into (a) the WLMservice policy options by which such pain and pleasure can bedelivered, (b) how to tailor these options for the results one seeks, (c)actual examples that delivered unintended results, and (d) therecommended practical solutions for the discussed problems.

SESSION 604 FRIDAY 8:00 AM - 9:00 AM

PAPER 5167 TECHNICALMAINFRAME (Z/OS) WINDOWS NT/XP/ETC.

Speaking SOA and Web Services:

.NET and the MainframeDr. Bernard Domanski, The City University of New York/CSI

Highlighting the re-emerging mainframe, this paper explores deployinga Service-Oriented Architecture (SOA) in a Microsoft-centricenvironment with a variety of available products from different vendors.We also look at Microsoft’s Host Integration Server 2004 to deploy aService-Oriented Architecture (SOA), which uses both mainframe and.NET technologies.


PAPER 5041 HOW-TONETWORK NON-SPECIFIC

Managing J2EE Applications with

ARM 4.0 and a Tutorial of the Open-

Source ARM 4.0 Java and C SDKBill Furnas, HPCarl J. De Pasquale, ADPMarcus Thoss, tang-IT Consulting GmbH

A technical and how-to presentation in two parts: Part 1: Dr. Carl DePasquale will discuss the results of his research on an approach formanaging J2EE applications using ARM 4.0, which can support anadaptive Application Management Framework (AMF). Part 2: BillFurnas and Marcus Thoss present a tutorial of the ARM 4.0 Java andC open-source SDK. Topics are: SDK features and supportedplatforms, and how-to build, test, run sample programs, and submitbugs and enhancements.

Chair: Ralph L. Gifford, AIG Data Center Inc.


PAPER 5011 ADVANCEDFUNDAMENTALS/CORE COMPETENCY NON-SPECIFIC

Distributed Resource Reclamation:

Enterprise Shared ServersStuart Plotkin, MetLife

Summer 2004, our company brought together all of our CapacityPlanners under one roof. From this group, our team was asked to leada high priority effort that was chartered with finding unused resourcesand reclaiming them. These resources could then be earmarked forredeployment. This paper focuses on the first domain we choose tobegin with consisting of 36 multi-processsor shared servers. Allprojects had been rolled out to new hardware within the past couple ofyears. This paper will have plenty of good ideas that YOU will be ableto use at your shop to do a lot with a little.


PAPER 5138 INTRODUCTORYWINDOWS NON-SPECIFIC

Monitoring the Monitors (If a Tree

Falls in a Forest, Will it Affect the

Network Bandwidth)Chris Greco, Social Security Administration

This presentation will show the challenge of attempting to consolidatealmost 30 monitoring tools into one cogent and predictive tool usingoff-the-shelf software products. The idea here is to show not a successbut a process through which companies are contacted, interviewedand in some cases provided a demo to try and establish a link betweensomewhat disparate troubleshooting solutions. It will show what isneeded is a standardization of the database; the framework of everymonitoring tool. Once databases are standardized, the overallmonitoring of these tools, will be realized.

Chair: Sidney W. Soberman, H W Wilson



DDF Performance Analysis - Does it

Really Have to be This Complicated?Robert E. Chaney, Delta Technology, Inc.

DB2 DDF processing, where the database is centralized and theapplication is distributed, offers some interesting challenges whendiagnosing ‘’End-To-End’’ performance problems. This paper uses anumber of different sources (from SMF to UNIX log files to webreports), to create a single view of a distributed workload as part of aperformance problem diagnosis. Warning: Success is NOTguaranteed!

52

FRIDAY SESSIONS

DE

CE

MB

ER

• 4-9

FRIDAY SESSIONS


PAPER 5057 TECHNICALNETWORK NON-SPECIFIC

Performance Modeling and Analysis

of Web SwitchJie Lu, BMC SoftwareJie Wang, University of Massachusetts Lowell

Web switch is a key component of the Web cluster architecture. It isoften built on a network processor (NP). This paper presents ananalytic performance model for analyzing performance of NP-basedWeb switches. The model contains parameters of systemconfigurations, dispatching algorithms, and workload characterizations,allowing users to trouble shoot performance bottlenecks and provisionthe capacity of the Web switch.



A Methodology for Predicting the

Scalability of Distributed Production

SystemsDr. Charles A. Letner, ALLTEL Communications, Inc.Richard Gimarc, HyPerformix, Inc.

A methodology is presented for developing a capacity-planning modelfor a highly distributed production environment in which EnterpriseApplication Integration (EAI) is a significant architectural element. Themethodology is demonstrated by building and validating a model usingdata collected from a production system. The level of abstraction isdescribed, along with a variation on the traditional model validationtechnique. The results demonstrate the usefulness of the methodologyand the resulting models for predicting hardware scalability in highlydistributed production environments.

SESSION 623 FRIDAY 10:30 AM - 12:00 PM


Making Your Web Portal a Dynamic

WebsiteFrank Bereznay, Kaiser Permanente

This paper examines the evolution of one reporting portal from a staticto a dynamic mode of operation. The benefits of a user-customizable,data driven, webpage authoring capability will be demonstrated usingASP to create dynamic HTML and a z/OS HFS to store large quantitiesof reporting objects created from an MXG PDB. The distinctionbetween dynamic HTML and dynamic report generation will beexplored along with a look at some SAS and non-SAS tools that can beused to dynamically create reporting objects from a SAS data source inresponse to user requests.


PAPER 5168 ADVANCEDMAINFRAME (Z/OS) Z/OS

LSPR Benchmark ConverterCharles E. Hackett, IBM

A set of trained neural networks are used as mapping functions totranslate the published LSPR benchmarks between the zOS 1.4 andOS/390 v2.10 sets. The neural networks provide an estimate of thelatest zSeries performance under the old benchmarks. Capacityplanners using previously validated LSPR workload distributions canprovide an estimate directly comparable to the older machines beingreplaced.



It’s All About Statistics!Rupa S. Joshi, Bank Of America

How can statistical analysis bring new insight to process management?Six Sigma, which means six standard deviations from the arithmeticmean, is a measurement-based methodology for eliminating defects, orerrors from any process. This paper is an introduction to the statisticscommonly used in Six Sigma’s DMAIC phases. A case study highlightsthe use of statistical tests and the appropriate use of confidenceintervals, P value and control charts to validate the data and determineits statistical significance. Finally, a road map to select the appropriatetests is discussed and outlined.

53

DE

CE

MB

ER

• 4-9

SESSION CROSS REFERENCE

55

SESSION CROSSREFERENCE

SUBJECT AREAS

• FUNDAMENTALS / CORE COMPETENCY

• HOT TOPICS

• ITIL

• MAINFRAME (Z/OS)

• NETWORK / INTERNET

• STORAGE

• UNIX / LINUX

• WINDOWS

PLATFORMS

• NON-SPECIFIC

• UNIX / LINUX

• WINDOWS NT / XP ETC.

• Z/OS

PLAPLATFORMSTFORMS

*nix = UNIX / LINUXNon = Non-Specific

Win = Windows NT/XP/etc.z/OS = z/OS

56

PLATFORM SESSION AUTHOR TITLE

*nix 335 Shanti Subramanyam Performance Management of a J2EE Application to Meet Service Level Agreements

*nix 345 Peter Johnson Java Garbage Collection Performance Analysis 201

*nix 447 Chandra Lanka Capacity Planning for Shared Middleware Environments; A Methodology

*nix 616 Dr. Charles A. Letner A Methodology for Predicting the Scalability of Distributed Production Systems

Non 301 K. Mani Chandy Sense and Respond Systems

Non 317 Dr. James Bouhana Analysis of Workload Alerts in Consolidated Servers

Non 327 Dr. Thomas E. Bell Determining Architectures of Existing Systems

Non 331 Henry H. Liu Service Demand Models for Enterprise Software Applications

Non 337 Michael Wiener Calculating Expected Reliability of Systems and Hardware

Non 341 Michael D. Maddox Using Fuzzy Logic to Automate Performance Analyses

Non 344 Prof. Ethan Bolker Virtual Performance Won’t Do: Capacity Planning for Virtual Systems

Non 416 Alexander Podelko Workload Generation: Does One Approach Fit All?

Non 425 Jack B. Woolley Want to Know WHY Response Time is So Long? Listen to the Wire.

Non 426 G Jay Lipovich Is It Time for Capacity Planners to Hang Up Their Cleats?

Non 427 Kevin Mobley Using Six Sigma to Define the Focus of Software Performance Engineering

Non 431 Dr. Jeffrey P. Buzen Workload Characterization for Parallel Processing Environments

Non 437 Irving Smith Wholesale Distributed Capacity Planning

Non 441 Dr. Jozo J. Dujmovic QNS - an Online System for the Study of Queuing Models

Non 506 Dr. Pierre M. Fiorini Software Performance Engineering Considerations in Unreliable Computing Environments

Non 507 Denise P. Kalm Survivor - The Corporate Jungle

Non 514 Linwood Merritt Closing the Gaps – Understanding Capacity Summarization

Non 517 James A. Yaple Benchmarking 101

Non 526 Dr. Mary R. Hesselgrave Underground SPE: Moving from Performance QA to SPE

Non 526 Heidi A. Gilmore Choosing a Load Testing Strategy

Non 547 Dr. Anatoliy Rikun Using Principal Component Method for Performance Data Compression and Analysis

Non 606 Stuart Plotkin Distributed Resource Reclamation: Enterprise Shared Servers

Non 623 Frank Bereznay Making Your Web Portal a Dynamic Website

Non 626 Rupa S. Joshi It’s All About Statistics!

Win 315 Peter Johnson Scaling up the JBoss Application Server

Win 324 Gene P. Fernando To V or not to V: A Practical Guide to Virtualization

Win 413 Peter Sevcik The Application Performance Index (Apdex) Standard+

Win 512 Ellen M. Friedman Measuring Performance in the Lab and Validating it in Production

z/OS 311 Dr. Lloyd G. Williams QSEM: Quantitative Scalability Evaluation Method

z/OS 417 Dick Arnold Reactive Capacity Planning – An Alternative

316 TBD Are Integrated Management Solutions the Recipe for an Invisible and Secure IT Infrastructure?

322 TBD UKCMG Best Paper:

326 Glenn O’Donnell Making Sense of the Performance Riddle

436 TBD CMG Australia Best Paper:

542 TBD CMG Italia Best Paper:

FUNDAMENTALS / CORE COMPETENCY

SUBJECT AREA CROSS REFERENCE


Non 314 Daniel A. Menasce Virtualization: Concepts, Applications, and Performance Modeling

Non 334 Peter J. Weilnau Measuring Up for Server Virtualization

Non 347 Chris Molloy Introduction to Data Center Markup Language (DCML)

Non 411 Dr. Jeffrey P. Buzen SOA and the Social Order - City Planning, Post Office & Business Protocols

Non 446 Dr. Annie W. Shum Through the Prism of Fractals: Why SOA Should Reflect the Natural Order

Non 504 Cathy Nolan Encryption Primer: An Introduction to Data Protection

Non 513 Marina Cismas Redefining Capacity Planning for Grid Computing

Non 516 Dr. Connie U. Smith A Performance Model Web Service

Non 543 David R. Morley CMG: The Early Years

Win 546 William L. Shelden, Jr., Ph.D. Modeling VMware ESX Performance

325 Dr. Toufic Boubez Policy Driven SOA

336 Herb VanHook Opportunities and Challenges of the SOA World

346 Eric Pulier Learning to Play Nicely

415 Nick Gall Reflection: The Next Big Thing After SOA

435 Ken Traub Radio-Frequency Identification at Enterprise Scale

536 Richard Soley The Model-Driven (R)evolution

HOT TOPICS


MAINFRAME (Z/OS)


57

Win 604 Dr. Bernard Domanski Speaking SOA and Web Services: .NET and the Mainframe

z/OS 312 Dr. Annie W. Shum A Multi-Tiered Approach With Data Normalization To Analyzing CPU Metrics

z/OS 322 Kenneth D. Williams MVS Application Performance Management

z/OS 332 Craig Hodgins Java Performance on z/OS: A Report from the Front Lines

z/OS 342 Don Deese Introduction to zSeries Application Assist Processor (zAAP)

z/OS 412 Robert R. Rogers How Do You Do What You Do When You’re a CPU?

z/OS 422 Glenn R. Anderson WebSphere Application Server for z/OS Version 6 - Performance Tuning

z/OS 432 Tony Ruberry zSeries Capacity Management - a True Story

z/OS 442 Ian C. Baldwin Migrating to z-990 - A User Experience

z/OS 501 Cheryl Watson z990 Performance and Capacity Planning Issues

z/OS 502 Stephen R. Guendert Taking FICON to the Next Level-Cascaded High Performance FICON

z/OS 511 Ned A. Diehl DB2 CPU and Response Metrics

z/OS 521 Kathy Walsh The Mechanics of Developing a High Quality Capacity Plan

z/OS 522 Thomas A. Halinski Unveiling of DB2’s DDF: SQL Revealed via the Gestalt Perspective

z/OS 531 Peter Enrico WLM Sysplex Management

z/OS 537 Michael P. Swanson 8 Great Myths of Software Asset Management

z/OS 541 Ivan L. Gelb Panel: zSeries Performance Q & A

z/OS 603 Ivan L. Gelb WLM Caused Pain and Pleasure

z/OS 614 Robert E. Chaney DDF Performance Analysis - Does it Really Have to be This Complicated?

z/OS 624 Charles E. Hackett LSPR Benchmark Converter


Non 313 Chris Molloy ITIL Capacity Management Deep Dive

Non 327 Avtar Dhillon Achieve IT Agility by Integrating SOA with ITIL Based BSM

ITIL

58


*nix 323 Bruce Naegel SMI (SNIA) Performance Update

*nix 434 Dan W. Yee I/O Performance Characteristics for Volume Managers on Linux 2.6 Servers

*nix 534 Prof. Robert Geist Enhancing Web Server Performance Through the Use of a Drop-In, Statically Optimal Disk

Scheduler

Non 333 Mel Boksenbaum Storage Performance Council Panel Discussion

Non 343 Bruce McNutt Disk Arm Management of Competing Workloads

Non 414 Charles T. McGavin Jr. Smoke and Mirrors – A Survey of Remote Replication Technologies

Non 424 Kathleen N. Hodge Database Disk to Disk Backups Using ATA Disk

Non 444 Robert Rogers Information Classification and Service Level Objectives for Information Lifecycle Management

Non 544 Chao Li Analytic Way for Performance Management of SAN

Win 424 Michael A. Salsburg A Management Framework For Petabyte-Scale Disk Storage

z/OS 524 Stephen R. Guendert Proper Sizing and Modeling of ESCON to FICON Migrations

STORAGE


*nix 445 Prof. James Westall Performance Tuning of Gigabit Network Interfaces

Non 425 Neil Carter Intrusion Detection/Prevention Devices — Are They Protecting Your Network - or Hampering It?

Non 545 Dr. Swami Ramany Workload modeling of Stateful Protocols Using HMMs

Non 605 Bill Furnas Managing J2EE Applications with ARM 4.0 and a Tutorial of the Open-Source ARM 4.0 Java and

C SDK

Non 615 Jie Lu Performance Modeling and Analysis of Web Switch

z/OS 525 Nalini J. Elkins One-Minute TCP Stack Analysis

505 Bernie Davidovics Measuring End User Response Time - With a Passive Network Probe

515 Laura Knapp Introduction to Simple Network Management Protocol (SNMP) Version 3

535 Chris Loosely Measuring Actual and Perceived User Experience on the Web

NETWORK / INTERNET


59



*nix 321 Mark M. Maccabee Testing Scalability of a WebLogic Application


*nix 523 Richard K. Roehl Designing a Fairer Round Robin Scheduling Algorithm



*nix 527 Igor A. Trubin Capturing Workload Pathology by Statistical Exception Detection System


Non 532 Rick Lebsack Facilitated Discussion: Future of the Performance Field 2005

z/OS 321 William R. Sullivan Overcoming Limitations to Java Application Scalability

UNIX / LINUX

PLAPLATFORMSTFORMS

*nix = UNIX / LINUXNon = Non-Specific

Win = Windows NT/XP/etc.z/OS = z/OS


Non 613 Chris Greco Monitoring the Monitors (If a Tree Falls in a Forest, Will it Affect the Network Bandwidth)

Win 423 Mark B. Friedman Virtual Memory Constraints in 32-bit Windows: an Update


Win 443 Michael A. Salsburg, PhD Panel: IT’s Role in Business Performance Management

Win 503 Jeffry A. Schwartz Understanding and Interpreting SQL Server Performance Counter

401 Amy D. Wohl The China Era

WINDOWS

PLATFORM CROSS REFERENCE

60


*nix 532 Rick Lebsack Facilitated Discussion: Future of the Performance Field 2005

Core 301 K. Mani Chandy Sense and Respond Systems

Core 317 Dr. James Bouhana Analysis of Workload Alerts in Consolidated Servers

Core 327 Dr. Thomas E. Bell Determining Architectures of Existing Systems

Core 331 Henry H. Liu Service Demand Models for Enterprise Software Applications

Core 337 Michael Wiener Calculating Expected Reliability of Systems and Hardware

Core 341 Michael D. Maddox Using Fuzzy Logic to Automate Performance Analyses

Core 344 Prof. Ethan Bolker Virtual Performance Won’t Do: Capacity Planning for Virtual Systems

Core 416 Alexander Podelko Workload Generation: Does One Approach Fit All?

Core 425 Jack B. Woolley Want to Know WHY Response Time is So Long? Listen to the Wire.

Core 426 G Jay Lipovich Is It Time for Capacity Planners to Hang Up Their Cleats?

Core 427 Kevin Mobley Using Six Sigma to Define the Focus of Software Performance Engineering


Core 437 Irving Smith Wholesale Distributed Capacity Planning

Core 441 Dr. Jozo J. Dujmovic QNS - an Online System for the Study of Queuing Models

Core 506 Dr. Pierre M. Fiorini Software Performance Engineering Considerations in Unreliable Computing Environments

Core 507 Denise P. Kalm Survivor - The Corporate Jungle

Core 514 Linwood Merritt Closing the Gaps – Understanding Capacity Summarization

Core 517 James A. Yaple Benchmarking 101

Core 526 Dr. Mary R. Hesselgrave Underground SPE: Moving from Performance QA to SPE

Core 526 Heidi A. Gilmore Choosing a Load Testing Strategy

Core 547 Dr. Anatoliy Rikun Using Principal Component Method for Performance Data Compression and Analysis

Core 606 Stuart Plotkin Distributed Resource Reclamation: Enterprise Shared Servers

Core 623 Frank Bereznay Making Your Web Portal a Dynamic Website

Core 626 Rupa S. Joshi It’s All About Statistics!

Hot 314 Daniel A. Menasce Virtualization: Concepts, Applications, and Performance Modeling

Hot 334 Peter J. Weilnau Measuring Up for Server Virtualization

Hot 347 Chris Molloy Introduction to Data Center Markup Language (DCML)

Hot 411 Dr. Jeffrey P. Buzen SOA and the Social Order - City Planning, Post Office & Business Protocols

Hot 446 Dr. Annie W. Shum Through the Prism of Fractals: Why SOA Should Reflect the Natural Order

Hot 504 Cathy Nolan Encryption Primer: An Introduction to Data Protection

Hot 513 Marina Cismas Redefining Capacity Planning for Grid Computing

Hot 516 Dr. Connie U. Smith A Performance Model Web Service


ITIL 313 Chris Molloy ITIL Capacity Management Deep Dive

ITIL 327 Avtar Dhillon Achieve IT Agility by Integrating SOA with ITIL Based BSM

Net 425 Neil Carter Intrusion Detection/Prevention Devices — Are They Protecting Your Network - or Hampering It?

Net 545 Dr. Swami Ramany Workload modeling of Stateful Protocols Using HMMs

Net 605 Bill Furnas Managing J2EE Applications with ARM 4.0 and a Tutorial of the Open-Source ARM 4.0 Java and C SDK

Net 615 Jie Lu Performance Modeling and Analysis of Web Switch

Stor 333 Mel Boksenbaum Storage Performance Council Panel Discussion

Stor 343 Bruce McNutt Disk Arm Management of Competing Workloads

Stor 414 Charles T. McGavin Jr. Smoke and Mirrors – A Survey of Remote Replication Technologies

Stor 424 Kathleen N. Hodge Database Disk to Disk Backups Using ATA Disk

Stor 444 Robert Rogers Information Classification and Service Level Objectives for Information Lifecycle Management

Stor 544 Chao Li Analytic Way for Performance Management of SAN

Win 613 Chris Greco Monitoring the Monitors (If a Tree Falls in a Forest, Will it Affect the Network Bandwidth)

NON-SPECIFIC

61



UNIX / LINUX

*nix 321 Mark M. Maccabee Testing Scalability of a WebLogic Application


*nix 523 Richard K. Roehl Designing a Fairer Round Robin Scheduling Algorithm



*nix 527 Igor A. Trubin Capturing Workload Pathology by Statistical Exception Detection System


Core 335 Shanti Subramanyam Performance Management of a J2EE Application to Meet Service Level Agreements

Core 345 Peter Johnson Java Garbage Collection Performance Analysis 201

Core 447 Chandra Lanka Capacity Planning for Shared Middleware Environments; A Methodology

Core 616 Dr. Charles A. Letner A Methodology for Predicting the Scalability of Distributed Production Systems

Net 445 Prof. James Westall Performance Tuning of Gigabit Network Interfaces

Stor 323 Bruce Naegel SMI (SNIA) Performance Update

Stor 434 Dan W. Yee I/O Performance Characteristics for Volume Managers on Linux 2.6 Servers

Stor 534 Prof. Robert Geist Enhancing Web Server Performance Through the Use of a Drop-In, Statically Optimal Disk Scheduler





WINDOWS NT / XP ETC.

Core 315 Peter Johnson Scaling up the JBoss Application Server

Core 324 Gene P. Fernando To V or not to V: A Practical Guide to Virtualization

Core 413 Peter Sevcik The Application Performance Index (Apdex) Standard+

Core 512 Ellen M. Friedman Measuring Performance in the Lab and Validating it in Production

Hot 546 William L. Shelden, Jr., Ph.D. Modeling VMware ESX Performance

Stor 424 Michael A. Salsburg A Management Framework For Petabyte-Scale Disk Storage

Win 423 Mark B. Friedman Virtual Memory Constraints in 32-bit Windows: an Update


Win 443 Michael A. Salsburg, PhD Panel: IT’s Role in Business Performance Management

Win 503 Jeffry A. Schwartz Understanding and Interpreting SQL Server Performance Counters

zOS 604 Dr. Bernard Domanski Speaking SOA and Web Services: .NET and the Mainframe

62






*nix 321 William R. Sullivan Overcoming Limitations to Java Application Scalability

Core 311 Dr. Lloyd G. Williams QSEM: Quantitative Scalability Evaluation Method

Core 417 Dick Arnold Reactive Capacity Planning – An Alternative

Net 525 Nalini J. Elkins One-Minute TCP Stack Analysis

Stor 524 Stephen R. Guendert Proper Sizing and Modeling of ESCON to FICON Migrations

zOS 312 Dr. Annie W. Shum A Multi-Tiered Approach With Data Normalization To Analyzing CPU Metrics

zOS 322 Kenneth D. Williams MVS Application Performance Management

zOS 332 Craig Hodgins Java Performance on z/OS: A Report from the Front Lines

zOS 342 Don Deese Introduction to zSeries Application Assist Processor (zAAP)

zOS 412 Robert R. Rogers How Do You Do What You Do When You’re a CPU?

zOS 422 Glenn R. Anderson WebSphere Application Server for z/OS Version 6 - Performance Tuning

zOS 432 Tony Ruberry zSeries Capacity Management - a True Story

zOS 442 Ian C. Baldwin Migrating to z-990 - A User Experience

zOS 501 Cheryl Watson z990 Performance and Capacity Planning Issues

zOS 502 Stephen R. Guendert Taking FICON to the Next Level-Cascaded High Performance FICON

zOS 511 Ned A. Diehl DB2 CPU and Response Metrics


zOS 522 Thomas A. Halinski Unveiling of DB2’s DDF: SQL Revealed via the Gestalt Perspective

zOS 531 Peter Enrico WLM Sysplex Management

zOS 537 Michael P. Swanson 8 Great Myths of Software Asset Management

zOS 541 Ivan L. Gelb Panel: zSeries Performance Q & A

zOS 603 Ivan L. Gelb WLM Caused Pain and Pleasure

zOS 614 Robert E. Chaney DDF Performance Analysis - Does it Really Have to be This Complicated?

zOS 624 Charles E. Hackett LSPR Benchmark Converter

Z/OS

DE

CE

MB

ER

• 4-9

EXHIBITOR DESCRIPTIONS

63


• EXHIBITOR INFORMATION

• EXHIBIT HALL HOURS *

Tuesday 11:30 AM – 4:00 PM

Wednesday 11:30 AM – 4:00 PM

Thursday 11:30 AM – 2:00 PM

* Hours are subject to change

CM

G2

00

5•

Or

la

nd

oEXHIBITOR DESCRIPTIONS

64

AESAES 149 Commonwealth Drive Booth: 500Menlo Park, CA 94025 Phone: 650-617-2400www.aesclever.com Fax: 650-617-2420

AES CLEVER® Solutions (CleverView for RoutePerformance, CleverView for IP Service Performance,CleverView for TCP/IP, CleverView for TCP/IP SLR,CleverView for Web Performance, and CleverView forcTrace Analysis) extend traditional network and Web sitemanagement capabilities into all corners of the IP networkworld. These proactive network performance and availabilityBusiness Service Management solutions focus on maximizingthe efficiency of critical TCP/IP network and Websiteresources and provide intelligent views into the world of IPnetworks and services.

Aptitune CorporationPO Box 1033 Booth: 518Plaistow, NH 03865 Phone: 603-382-4200www.sarcheck.com Fax: 603-382-4247

Aptitune produces performance analysis tools for UNIX andLinux, including our flagship product, SarCheck®. SarCheckanalyzes sar reports, data from the /proc filesystem and ps -elf output. Tunable kernel parameters are examined toproduce a concise plain English report. SarCheck identifiesproblems and recommends corrective action. SarCheck canuse gnuplot to create JPEG or PNG graphs accompanying itstext.

We’re the folks with the propeller hats! – come by and get onewhile they last.

ASG1333 Third Avenue South Booth: 307Naples, FL 34102 Phone: 239-435-2200www.asg.com

Founded in 1986, ASG is a privately held enterprise softwareand professional services firm that provides a full range ofsoftware solutions in the Metadata Management, ApplicationsManagement, Operations Management, ContentManagement, Performance Management, SecurityManagement, and Infrastructure Management arenas. ASG isheadquartered in Naples, Florida, USA, with more than 45offices throughout the Americas, Europe, Middle East, Africa,and Asia/Pacific. For more information, visit ASG atwww.asg.com.

BMC Software2101 CityWest Blvd Booth: 301Houston, Tx 77042 Phone: 713-918-8800www.bmc.com

BMC Software, Inc., is a leading provider of enterprisemanagement solutions that empower companies to managetheir IT infrastructure from a business perspective. DeliveringBusiness Service Management, BMC Software solutionsspan enterprise systems, applications, databases and servicemanagement. Founded in 1980, BMC Software has officesworldwide and fiscal 2004 revenues of more than $1.4 billion.For more information about BMC Software, visitwww.bmc.com.

CapTell Developments Pty LtdP.O. Box 3239 Manuka Booth: 204Canberra, ACT 2603 Phone: +61 (0)2 6232 6364www.captelldevelopments.com Fax: +61 (0)2 6232 6365

CapTell Developments, an Australian company provides costeffective capacity management reporting solutions. CapTell isan easy to use software tool for the interpretation of computerusage data and the production of meaningful capacitymanagement reports. It allows users to tailor documenttemplates and includes embedded objects such as graphsand tables. Once the templates have been established, userscan easily create standard reports with the latest dataavailable. With CapTell, cyclic reporting can be fullyautomated freeing up the capacity manager for moreappropriate analysis work.

CIMS Lab, Inc.3013 Douglas BLVD Suite 120 Booth: 419Roseville, CA 95661 Phone: 916-783-8525www.cimslab.com Fax: 916-783-2090

The CIMS Suite of products provides Capacity Planning,Resource Accounting and Chargeback across the entireEnterprise. CIMS data collectors support all major platformsand systems such as UNIX/Linux, Windows, Mainframe,Databases, E-Mail Systems, Networks, Internet and more.

CIMS is always working with its customers to develop newfeatures and products to improve Capacity Planning,Resource Accounting and Chargeback,

To learn more about the CIMS system, please stop by theCIMS booth (#419).

DE

CE

MB

ER

• 4-9

Computer Management Sciences, Inc.6076-D Franconia Road Booth: 416Alexandria, VA 22310 Phone: (703) 922-7027www.cpexpert.com Fax: (703) 922-7305

CMS offers CPExpert to help analyze performanceproblems with MVS environments. CPExpert automatescomputer performance evaluation by detecting systemperformance problems, identifying the causes of theproblems, & suggesting ways to improve systemperformance. Different components of CPExpert analyzeperformance problems with WebSphere MQ, DB2, GoalMode, CICS, & DASD.

Demand Technology Software, Inc.1020 8th Ave. S. #6 Booth: 509Naples, FL 34102 Phone: 239-261-8945www.demandtech.com Fax: 239-261-5456

Demand Technology Software is the proud developer ofPerformance SeNTry, the premiere Windows Server 2003performance data collector. Performance SeNTry is designedwith the professional performance analyst and capacityplanner in mind. When you combine Performance SeNTry’sintelligent data collection capabilities; its full function MicrosoftSQL Server-based Performance DataBase; and third partysupport from Lund MetaView and Performance Gallery,Merrill’s MXG, SAS, and SAS Institute’s IT ResourceManager, you have the ability, to take control of your criticalWindows servers and applications.

Diversified Software18635 Sutter Blvd. Booth: 312Morgan Hill, CA 90537 Phone: 408-778-9914www.diversifiedsoftware.com Fax: 408-778-6875

Diversified Software is the leading supplier of JCLmanagement products to data centers throughout the world.Over 1,000 of the top 2,000 data centers worldwide rely onone or more of the company’s products. Founded in 1978,Diversified Software continues its commitment to thechallenging world of JCL management through extensiveresearch and development and by providing the criticaltechnical support needed in today’s complex IT environments.Diversified Software products are available directly from thecompany and through the IBM remarketing program.

HyPerformix, Inc.4301 Westbank Drive Bldg. A, Suite 300 Booth: 512Austin, TX 78746 Phone: 512-328-5544www.hyperformix.com Fax: 866-495-4293

HyPerformix is a leading application performance andcapacity management software company focused on theneeds of Global 2000 enterprises who share three commongoals:Deliver applications that meet performance objectives;Maximize IT investments; and Manage the risk of constantchange. HyPerformix helps reach these goals through aproactive approach to performance and capacitymanagement that provides a service level view across theentire application life cycle. This end-to-end view of theapplication ensures alignment between IT resources andbusiness needs.

ISM (the Information Systems Manager)One Bethlehem Plaza Booth: 507Bethlehem, PA 18018 Phone: 610-865-0300www.perfman.com Fax: 610-868-6277

ISM helps some of the world’s most successful companiesachieve extraordinary levels of quality, efficiency anddependability in Information Technology.

‘’ISM - The power behind great IT decisions.’’

itSMF USA1012 Porter Lane Booth: 201Normal, IL 61761 Phone: 309-826-4598www.itsmfusa.org Fax: 309-888-4869

Since our beginnings in 1997, the itSMF USA has grown tobecome the premier IT Service Management professionalforum in the USA. We have continuously taken exciting stepsin professionalizing and expanding our capabilities andservices for our membership. As a result, 2005 promises tobe our most exciting year yet! We have 29 active itSMF USALocal Interest Groups (LIGs) and growing.

Metron-Athene, Inc.6320 Canoga Avenue Suite 1500 Booth: 410Woodland Hills, CA 91367 Phone: 818-227-5019www.metron-athene.com Fax: 818-227-5099

Metron-Athene is a leading Performance Management,Capacity Planning and Capacity Forecasting, softwarespecialist. Athene, our flagship product is a widely adoptedsolution providing Performance Alerting, Automatic Reporting,Workload Characterization, Service Level Reporting, TrendAnalysis and Server Consolidation. Athene supports UNIX,Linux, Windows and z/OS mainframes.


65

CMG2005 BRONZE SPONSOR


CM

G2

00

5•

Or

la

nd

o

66

MVS Solutions Inc.8300 Woodbine Ave 4th Floor Booth: 300Markham, ON L3R9Y7 Phone: 905-940-9404www.mvssol.com Fax: 905-940-5308

If you’ve been given the task of improving your batch windowprocessing, you must look at ThruPut Manager. ThruPutManager 6 Automation Edition automates your batchprocessing, saving a great deal of elapsed processing time,shortening your batch window time and getting your reportsout and your online up faster. If you’re a CA-7 user, there’seven more! Learn how we can empower CA-7 and help it getyour schedule complete, on-time, every time. Look for ourvendor presentation and visit our booth to learn how ThruPutManager solves batch problems encountered by z/OS datacenters everywhere.

OPNET Technologies7255 Woodmont Ave. Booth: 406Bethesda, MD 20814 Phone: 240.497.3000www.opnet.com Fax: 240.497.3001

OPNET is a leading provider of management software fornetworks and applications, enabling enterprises to:

1. rapidly diagnose end-to-end performance problems, pre- or post-deployment

2. model server, mainframe, and network infrastructure for capacity planning

3. automate network diagnostics, for policy and configuration change validation

OPNET products deliver rapid ROI to thousands ofcustomers, including blue-chip organizations such as: AbbottLabs, Continental Airlines, Mary Kay, Office Depot, Prudential,TXU, and SunTrust Banks.

PerfCap Corporation85 Perimeter Road Booth: 517Nashua, NH 03063 Phone: 603-594-0222www.PerfCap.com Fax: 60-594-0826

PerfCap offers over 20 years of expertise in developing out-of-the-box, fully automated, web-based, highly scalable andeconomical software for Performance, Capacity and AssetManagement. PAWZ: Performance Software for historical andreal-time data collection, analysis, reporting and alerting.eCAP: A Capacity Planning software to answer “What if”questions using analytical queuing network modeling. FindIT:An Asset Management and Change Tracking Softwareproviding inventory of hardware, software and licenses.Platforms: HP-UX, Tru64, Sun-Solaris, IBM-AIX, Linux,OpenVMS, Windows NT/2000/2003.

ProactiveNet, Inc.2055 Laurelwood Road Suite 130 Booth: 505Santa Clara, CA 95054 Phone: 408-454-4500www.proactivenet.com Fax: 408-454-4501

ProactiveNet, Inc. is the time-to-value leader in BusinessServices Management software solutions for Global 2000 andleading e-commerce companies. The company’s end-to-endsolutions deliver unique capabilities to customers withrequirements to align IT services to Business priorities,seamlessly link Business process performance to IT serviceand infrastructure behaviors, and maximize IT’s value to thebusiness.

Responsive Systems281 Hwy 79 Booth: 401Morganville, NJ 07751 Phone: 732-972-1261www.responsivesystems.com

Responsive Systems provides the proven products for tuningyour DB2 and IMS systems. Buffer Pool Tool ® for DB2 andIMS. Dynamic SQL Cache Control. Many of our client sites

have saved millions in the first year of use.

SAS Institute Inc.100 SAS Campus Drive Booth: 207Cary, NC 27513 Phone: 919-677-8000www.sas.com Fax: 919-677-4444

SAS is the market leader in providing a new generation ofbusiness intelligence software and services that create trueenterprise intelligence. SAS solutions are used at more than40,000 sites - including 96 of the top 100 companies of theFORTUNE Global 500®. For nearly three decades, SAS hasbeen giving customers around the world The Power toKnow®.

Join us Monday, Dec. 5, to learn how SAS® IT ManagementSolutions help you manage your IT organization and servicedelivery, control costs and make informed decisions thatdirectly affect your bottom line.



DE

CE

MB

ER

• 4-9


67

Segue Software201 Spring Street N/A Booth: 410Lexington, MA 02421 Phone: 781.402.1000www.segue.com Fax: 781.402.1094

Segue Software is a global expert in delivering solutions todefine, measure, manage and improve software qualitythroughout the entire software application lifecycle. Segue’sSoftware Quality Optimization™ (SQO™) solutions helpcompanies reduce business risk, ensure the deployment ofhigh quality software and increase ROI. Businesses aroundthe world rely on Segue’s Silk products to protect businessservice levels, competitive edge and brand reputation.Headquartered in the U.S., with offices across North America,Europe and Asia, Segue can be reached at +1-781-402-1000or www.segue.com.

Symmetricom34 Tozer Road Booth: 213Beverly, MA 01915 Phone: 978-232-1477www.ntp-systems.com Fax: 978-927-4099

Symmetricom is the only supplier of comprehensive networktime synchronization solutions that include the dedicatednetwork time servers and the necessary synchronization,management and monitoring software that synchronizes thetime on IT devices such as workstations, servers and routers.Our solutions keep log files accurate, applications correct andoffer the capability to monitor and manage all of the clients onthe network.

TeamQuest CorporationOne TeamQuest Way Booth: 407Clear Lake, IA 50428 Phone: 641-357-2700www.teamquest.com Fax: 641-357-2778

TeamQuest Corporation is the global leader in IT serviceoptimization (ITSO). TeamQuest Performance Software helpscompanies quickly manage the efficiency and availability ofservices they deliver to client organizations. Specializing incapacity planning and performance management, TeamQuestCorporation helps ensure availability and reliability ofenterprise systems with advanced software solutions.TeamQuest is a trusted provider of software solutions thatimprove service delivery, proactively align business prioritieswith IT resources, reduce risk, and cut costs in their ITorganizations.

Trident Services, Inc.1260 41st Avenue, Suite K Booth: 413Capitola, CA 95010 Phone: 831-465-7661www.triserv.com Fax: 831-476-0966

Trident Services, Inc. has announced Version 6.0 for z/OS.OS/EM can help today’s Data Center to improvemanageability, improve system throughput, service levels,reduce hardware expenses, software license costs andreduce human resource expense. New functionality includesImproved throughput management by managing Data SetName Enqueue conflicts, Additional JCL selector criteria forbatch resource routing and job classing functions, EnhancedRACF password content controls, Dynamic STEPLIB for TSOREXX Execs, Limit Concurrent Execution of programs andHSM Data Set recalls prior to job execution.

Uptime Software555 Richmond Street West Booth: 408P.O. Box 110, Suite 408 Phone: 416-868-0152Toronto, ON M5V 3B1 Fax: 416-868-4867www.uptimesoftware.com

Uptime software is a leading innovator in server performanceand availability management. Our software solution providesclients in over 30 countries with enterprise server capacity andavailability monitoring, 7x24 via web browser, deliversautomated reports to allow rapid problem isolation, trendanalysis and simplified capacity planning. We help our clientstake control of their IT infrastructure, improve serviceavailability and optimize server capacity with an easy-to-usetool that costs less than competitive frameworks and deploysin minutes without formal training or consulting services.


CMG2005 PLATINUM SPONSOR

CMG2005 SILVER SPONSOR

Joan A. Arnold-RoksandichMel BoksenbaumClaire S. Cates

Michael B. MarcusTom MoulderCatherine S. Nolan

Richard S. RalstonSidney W. Soberman

Access ControlC

MG

20

05

• O

rla

nd

o

68

Joan A. Arnold-RoksandichJohn M. BaxterJoe E. BellThomas E. Bell PhDThomas BeretvasFrank M. BereznayShana J. BereznayJanet BishopMel BoksenbaumGerard J. BoninMartin D. BrakeGiuliano CasaleClaire S. CatesPablo ChacinJim CheesemanSteve ClarkThomas G. ConfreySteven CrowlFlorin DavidWilliam N. EagleJerrl M. EvansEllen M. FriedmanGary Y. FurukawaJohn GebhardtIvan L. GelbJonathan GladstoneBruce L. GreenMargaret R. GreenbergStephen GuendertMohammad Gulzar

Mary R. HesselgraveNorman HollanderCurtis Hrischuk, Ph.D.William J. HuebschMario JauvinMaryann JessenBrian M. JohnsonRobert A. JohnsonRupa S. JoshiWilliam JourisDenise P. KalmMary KarlinsMichael L. KnychMinda E. LarsonJaqueline A. LynchMichael B. MarcusStephen J. MarksamerKadhar MasthanCharles T. McGavin, Jr.Patricia L. McGuireMattie McKnightChris L. MolloyEdna B. MorrisonJoel MotznyTom MoulderAnthony G. MungalJohn MurphyChristian G. NielsenCatherine S. NolanTim R. Norton

Rich J. OlcottJohn R. PaulOdysseas I. Pentakalos, Ph.D.Bernard R. PierceMichael L. PuldyEdward RabinovitchRichard S. RalstonMichael S. RecantGerald N. RusthovenCharles SavageSusan SchreitmuellerJeffry A. SchwartzXianneng ShenJeffrey SmithBarry N. SokolikElizabeth A. SomersDaniel J. SquillaceKathy J. SteffensJules S. StollakShanti SubramanyamMike TarraniDavid C. ThornVictoria G. ToblemanPeter van EijkKenneth D. WilliamsBrian WilsonBernie K. WongJames Yaple

Editorial Review Board

ACKNOWLEDGEMENTS

DE

CE

MB

ER

• 4-9

ACKNOWLEDGEMENTS

69

Joan A. Arnold-RoksandichDenise ArrudaJohn BaxterJoe BellTom BellFrank BereznayShana BereznayRichard BergPaul BillickMel BoksenbaumGerry BoninGiuliano CasalePablo ChacinJim CheesemanMoses ChuSteve ClarkThomas ConfreySteven CrowlFlorin DavidKen DeVilbissPavel DournovWilliam N. EagleBill FairchildRosemary FlemingDonna FolkertsRex FrankEllen FriedmanMark FriedmanGary Y. FurukawaMichael GardnerIvan GelbBart GijsenJonathan GladstoneJoel GoldsteinMargaret GreenbergStephen GuendertAnkur Hajare

Warren HaywardMary HesselgraveJim HoltmanCurtis HrischukBill HuebschFrank J. IngrassiaBob JacobsMaryann JessenBrian JohnsonRobert JohnsonBill JourisDenise KalmMary KarlinsCraig KimGary KingGregory A. KorbeckiSamuel KounevSrini KrisnamurthyMinda LarsonRick LebsackHenry LiuJaqueline LynchMichael MarcusSteve MarksamerPatricia L. McGuireChristopher MellgrenLin MerrittSerg MescheryakovJane MessmerChris MolloyEdna MorrisonJoel MotznyTom MoulderAnthony MungalJohn MurphyCathy NolanTim R. Norton

Rich OlcottCharles PadamadanJohn PaulOdysseas PentakalosBernard PierceMichael PuldyEdward RabinovitchRick RalstonMichael RecantRusty RustovenMichael SalsburgSusan SchreitmuellerAlan SchulmanJeffry A. SchwartzXianneng ShenSean ShepleyJeffrey SmithSid SobermanBarry SokolikElizabeth SomersDan SquillaceKathy SteffensDamon StewartJules StollakShanti SubramanyamMike TarraniDave ThornRhane VennesRudy WaldnerYanyan WangRay WilliamsBrian WilsonBernie WongCathy WrightJames Yaple

Referees

CM

G2

00

5•

Or

la

nd

o

70

Joan A. Arnold-RoksandichThomas E. Bell PhDFrank M. BereznayShana J. BereznayPaul BillickMel BoksenbaumMartin D. BrakeClaire S. CatesBryan DrakeDonna S. FolkertsEllen M. FriedmanGary Y. FurukawaMichael GardnerIvan L. GelbRalph L. GiffordJonathan GladstoneJonathan GladstoneMargaret R. Greenberg

Stephen GuendertWarren HaywardJames HoltmanBrian M. JohnsonRupa S. JoshiAnnette S. KakazuDenise P. KalmMichael L. KnychGregory A. KorbeckiMinda E. LarsonMichael B. MarcusStephen J. MarksamerKadhar MasthanMattie McKnightChristopher MellgrenChris L. MolloyEdna B. MorrisonTom Moulder

Catherine S. NolanCharles PadamadanRichard S. RalstonMichael S. RecantJerry L. RosenbergRaymond R. RothNancy A. SabatiniSusan SchreitmuellerXianneng ShenSean ShepleySidney W. SobermanBarry N. SokolikElizabeth A. SomersBarry StantonBarry StantonDavid C. ThornRudy J. WaldnerJames Yaple

Session Chairs

Joan A. Arnold-RoksandichDenise T. ArrudaFrank M. BereznayShana J. BereznayJanet BishopMel BoksenbaumGerard J. BoninSteve ClarkEllen M. FriedmanGary Y. FurukawaMichael GardnerJonathan Gladstone

Bruce L. GreenMargaret R. GreenbergJames HoltmanBrian M. JohnsonRupa S. JoshiChandra LankaStephen J. MarksamerKadhar MasthanMattie McKnightEdna B. MorrisonTom MoulderCatherine S. Nolan

Charles PadamadanChris PaluziRichard S. RalstonMichael S. RecantRaymond R. RothNancy A. SabatiniSusan SchreitmuellerSean ShepleySidney W. SobermanBarry N. SokolikBarry StantonDamon Stewart

Session Monitors

Joseph L. BabcockJoe E. BellThomas E. Bell PhDFrank M. BereznayMartin D. BrakeClaire S. CatesEllen M. FriedmanGary Y. FurukawaIvan L. GelbMargaret R. GreenbergJames HoltmanMario Jauvin

Brian M. JohnsonWilliam JourisDenise P. KalmJaqueline A. LynchMichael D. MaddoxStephen J. MarksamerKadhar MasthanChris L. MolloyEdna B. MorrisonJoel MotznyAnthony G. MungalCatherine S. Nolan

Edward RabinovitchRichard S. RalstonSusan SchreitmuellerJeffry A. SchwartzBarry N. SokolikDaniel J. SquillaceBarry StantonElizabeth A. SomersDavid C. ThornKenneth D. WilliamsJames Yaple

Mentors

ACKNOWLEDGEMENTS

DE

CE

MB

ER

• 4-9

GENERAL INFORMATION

71

GENERAL INFORMATION

ATTENDEES WITH SPECIAL NEEDS

If you have special needs addressed by theAmericans with Disabilities Act, stop by theCMG2005 On-Site Registration Area. We will makeevery effort to accommodate your needs.

BADGE POLICY

To gain access to all CMG functions and meals,attendees must wear conference badges and have theappropriate ticket. There will be no exceptions.

Registered CMG guests must wear their badge andhave the appropriate ticket to gain access to breakfast,lunch and PARS events.

BIRDS-OF-A-FEATHER (BOF) SESSIONS

Have a specific interest you want to discuss?CMG2005 offers BOF sessions Tuesday, Wednesdayand Thurdays evenings from 6:15 PM - 7:15 PM.following sessions, these informal gatherings allowattendees with similar concerns to exchange ideas in asmall group setting.

CMG HEADQUARTERS HOURS

Saturday, December 3, 2005 4:00 PM – 8:00 PM

Sunday, December 4, 2005 7:00 AM – 12:00 PM

1:00 PM – 8:00 PM

Monday, December 5, 2005 7:00 AM – 12:00 PM

1:00 PM – 8:00 PM

Tuesday, December 6, 2005 7:00 AM – 12:00 PM

1:00 PM – 6:00 PM

Wednesday, December 7, 2005 7:00 AM – 1:30 PM

Thursday, December 8, 2005 7:00 AM – 1:30 PM

Friday, December 9, 2005 7:30 AM – 10:00 AM

EXHIBITOR DETAILS

The CMG2005 Exhibitor Hall will be located in Florida A.

EXHIBIT HALL HOURS




For more information on CMG2005 Exhibitors, refer tothe Exhibitor Section.

USER GROUP MEETINGS

Some CMG2005 exhibitors use Monday, December 5,to hold User Group Meetings.

EXHIBITOR PRESENTATIONS

Many of the CMG2005 exhibiting companies will holdexhibitor presentations during conference week.Scheduled from 5:00 PM – 6:00 PM on Tuesday,Wednesday and Thursday, these presentations allowattendees a more comprehensive look at the

exhibitor’s products and services.

CONFERENCE REGISTRATION INFORMATION

REGISTRATION HOURS

Saturday, December 3, 2005 4:00 PM – 8:00 PM

Sunday, December 4, 2005 7:00 AM – 12:00 PM

1:00 PM – 8:00 PM

Monday, December 5, 2005 7:00 AM – 12:00 PM

1:00 PM – 8:00 PM


1:00 PM – 6:00 PM



Friday, December 9, 2005 7:30 AM – 10:00 AM

CM

G2

00

5•

Or

la

nd

oMEAL SCHEDULES

Breakfast will be served Tuesday – Friday, inOsceola A

Lunch will be served Tuesday – Friday in Osceola A.Please note that lunch will be served promptly at12:00 PM.

Veggie/Kosher Meals – If you prefer vegetarian orkosher meals and did not indicate your preference onyour registration form, please notify a CMG staffmember at the On-Site Registration Area.

Individual meal tickets are also available at the CMGOn-Site Registration Area. Breakfast tickets are $20,lunch tickets are $25 and PARS tickets are $50 pernight.

MULLEN FOUNDATION

Extra CMG collectible items are available Wednesdayand Thursday. Times will be posted. All check or cashdonations will go directly to the J. William MullenFoundation. The Mullen Foundation cannot acceptcredit card contributions.

NEWSLETTER

Pick up your daily newsletter at Breakfast,Registration, or the Exhibitor Area for last minuteupdates to conference activities. A daily schedule ofBOF’s is found in each issue.

PARS

CMG hosts the unique Performance Analysts’Relaxation Session (PARS), Monday – Thursday, from7:30 PM – 11:00 PM Take a breather, network with otherattendees or just relax after a long day. Enjoy horsd’oeuvres during the first 90 minutes each evening,complimentary soft drinks, quiet areas, andentertainment. Please note: You must be 21 years of

age to attend PARS.

PUBLICATIONS FOR PURCHASE ONSITE

CMG has several member publications, in both printedand electronic formats. All attendees will receive theCMG2005 CD-ROM. Additional copies of theCMG2005 Proceedings (in either the 2-volume set oron CD-ROM) are available for purchase. For all othermember publications, please consult the publicationorder form found in the CMG2005 On-Site RegistrationArea. (Orders of older publications will be mailed inJanuary.)

RECRUITMENT POLICY

As the intent of the meeting is to share information, notto recruit new staff, there is a no recruiting policy atthe conference. Deliberate recruiting such as jobpostings, general announcements, or recruitingconversations with specific individuals, is specificallyprohibited at any meeting or activity sponsored byCMG.

SHIPPING INFORMATION

The hotel provides shipping services and a full serviceBusiness Center for your convenience.

SMOKING

CMG’s no smoking policy includes all CMG sessions,exhibit halls, meals, and PARS functions. There will bedesignated areas for smokers. We appreciate yourcompliance with this policy.

SUNDAY WORKSHOPS, DECEMBER 4, 2005

Breakfast 7:30 AM – 8:30 AM

Morning Workshops 8:30 AM – 12:00 PM

Lunch 12:00 PM – 1:00 PM

Afternoon Workshops 1:00 PM – 4:30 PM

You must be registered to attend Sunday Workshops.Each workshop will only be held once and will not berepeated. Attendees registered for the SundayWorkshops will receive meal tickets with theirworkshop materials for Sunday’s Breakfast and Lunch.

WHAT ARE THE CMG2005 COLLECTIBLES?

Your on-site registration packet contains your ticketsfor the conference collectibles, which include a FinalAgenda (on-site program), Proceedings and otherCMG gifts.

CONFERENCE GUEST REGISTRATION

Orlando is a fun place for everyone! In order to attendvarious CMG2005 functions, guests must be registeredby a CMG2005 conference attendee. Registrationforms are available in the CMG On-Site RegistrationArea. For only $350.00, your registered guest receivesa conference badge and can attend:

• PARS on Monday, Tuesday, Wednesday, andThursday evenings. A guest must have a badge andPARS ticket to attend PARS functions.

• Full Breakfast Tuesday – Friday in the designatedmeal area.

• Entrance into CMG’s Exhibit Hall on Thursday.

• Registered guests may attend only the session atwhich their sponsoring attendee makes apresentation. A guest registration does not includeattending formal conference sessions.

• Individual tickets for the luncheons or PARS can bepurchased at the CMG On-Site Registration Area.

72

GENERAL INFORMATION

DE

CE

MB

ER

• 4-9

GENERAL INFORMATION

73

CMG2005 PLATINUM SPONSOR

CMG2005 BRONZE SPONSORS

CMG WOULD LIKE TO THANK OUR SPONSORS

CMG2005 SILVER SPONSOR

C M G 2 0 0 5

COMPUTER MEASUREMENT GROUP

151 Fries Mill Road, Ste. 104Turnersville, NJ 08012

DECEMBER 4-9, 2005ORLANDO, FLORIDA

www.cmg.org/conference1

FEATURING:MAINFRAME (Z/OS)

UNIX / LINUX

STORAGE

FUNDAMENTALS / CORE COMPETENCY

NETWORK / INTERNET

STORAGE

WINDOWS

HOT TOPICS

ITIL

FIRST-CLASS PRSRTU.S. POSTAGE

PAIDPermit No. 2066Eau Claire, WI

cmg2005 • orlando · 2020-02-20 · cmg2005 • orlando ... seasoned veterans,...

Documents