Cyberinfrastructure and Internet2

Eric BoydDeputy Technology Officer

Internet2

What is Cyberinfrastructure (CI)?

• A strategic orientation supported by NSF

• Calls for large-scale public investment to encourage the evolution of widely distributed computing via the telecommunications network

• Goal is to deploy the combined capacity of multiple sites to support the advance of current research, initially in science and engineering

General Session Thursday: Cyberinfrastructure: The Way Forward

• Francine Berman, San Diego Supercomputer Center, Moderator

• Paul Avery, University of Florida

• Thomas Knab, Case Western Reserve University

• Alan Whitney, MIT Haystack Observatory

• Eric Boyd, Internet2

The Distributed CI Computer

Instrumentation

Security

Control

DataGeneration

Computation

Analysis

Simulation

Program

Security

ManagementSecurity and

AccessAuthentication

AccessControl

Authorization

Researcher

ControlProgram

ViewingSecurity

3DImaging

Display andVisualization

.

DisplayTools Security

DataInput

CollabTools Publishing

HumanSupportHelp

Desk

Policy andFunding

ResourceProviders

FundingAgencies

Campuses

SearchData SetsStorage

Security

RetrievalInput

SchemaMetadata

DataDirectories

Ontologies

Archive

EducationAnd

Outreach

Network

Training

The Network is the Backplane for the Distributed CI Computer

CollabTools Publishing

HumanSupportHelp

Desk EducationAnd

Outreach

Network

Training

Challenge and Opportunity

• Challenge:• The R&E community thinks of CI primarily in terms of building distributed computing clusters

• Opportunity:• The network is a key component of CI• Internet2 is leading the development of solutions for the network component of CI

CI Requirements

• Data storage• Robust campus infrastructure• Security and Authorization • IT support for local and remote resources

• Network performance monitoring tools• Network resources to meet demand spikes

7

LHC epitomizes the CI Challenge

Current Situation

• Large Hadron Collider (LHC) at CERN will go operational in 2008

• Over 68 U.S. Universities and National Laboratories are poised to receive data

• More than 1500 scientists are waiting for this data

• Are campus, regional, and national networks ready for the task?

9

CERNTier 0 Raw Data

10

CERNTier 0 Raw Data

FNAL BNL Shared Data Storage and Reduction

Tier 1(12 orgs)

US Tier 2(15 orgs)

CMS (7) Atlas (6-7)

Provides Data to Tier 3

US Tier 3 (68 orgs) Scientists Request

Data

US Tier 4 (1500 US scientists)

Scientists Analyze Data

11

CERNTier 0 Raw Data

FNAL BNL Shared Data Storage and Reduction

Tier 1(12 orgs)

US Tier 2(15 orgs)

CMS (7) Atlas (6-7)

US Tier 3 (68 orgs)

US Tier 4 (1500 US scientists)

Scientists Request Data

Provides Data to Tier 3

Scientists Analyze Data

LHCOPN

GEANT-ESNet-Internet2

Internet2/Connectors Internet2/Connectors

Local Infrastructure

12

CERN

Tier 0 to Tier1: Requires 10-40 Gbps

Tier 1 to Tier 2: Requires 10-20 Gbps

LHCOPN

GEANT-ESNet-Internet2

Internet2/Connectors Internet2/Connectors

Tier 1 or 2 to Tier 3: Estimate: Requires 1.6 Gbps per transfer (2 TB's in 3 hours)

Peak Flow Network Requirements

Local Infrastructure

13

Science Network Requirements Aggregation Summary(slide courtesy of ESNet)

Science Drivers

Science Areas / Facilities

End2End Reliability

Connectivity 2006 End2End Band width

2010 End2End Band width

Traffic Characteristics

Network Services

Advanced Light Source

- • DOE sites

• US Universities

• Industry

1 TB/day

300 Mbps

5 TB/day

1.5 Gbps

• Bulk data

• Remote control

• Guaranteed bandwidth

• PKI / Grid

Bioinformatics

- • DOE sites

• US Universities

625 Mbps

12.5 Gbps in two years

250 Gbps

• Bulk data

• Remote control

• Point-to-multipoint

• Guaranteed bandwidth

• High-speed multicast

Chemistry / Combustion

- • DOE sites

• US Universities

• Industry

- 10s of Gigabits per second

• Bulk data • Guaranteed bandwidth

• PKI / Grid

Climate Science

- • DOE sites

• US Universities

• International

- 5 PB per year

5 Gbps

• Bulk data

• Remote control

• Guaranteed bandwidth

• PKI / Grid

High Energy Physics (LHC)

99.95+%

(Less than 4 hrs/yea

r)

• US Tier1 (DOE)

• US Tier2 (Universities)

• International (Europe, Canada)

10 Gbps 60 to 80 Gbps

(30-40 Gbps per US Tier1)

• Bulk data

• Remote control

• Guaranteed bandwidth

• Traffic isolation

• PKI / Grid

Science Drivers

Science Areas / Facilities

End2End Reliability

Connectivity 2006 End2End Band width

2010 End2End Band width

Traffic Characteristics

Network Services

Magnetic Fusion Energy

99.999%

(Impossible

without full

redundancy)

• DOE sites

• US Universities

• Industry

200+ Mbps

1 Gbps • Bulk data

• Remote control

• Guaranteed bandwidth

• Guaranteed QoS

• Deadline scheduling

NERSC - • DOE sites

• US Universities

• Industry

• International

10 Gbps 20 to 40 Gbps

• Bulk data

• Remote control

• Guaranteed bandwidth

• Guaranteed QoS

• Deadline Scheduling

• PKI / GridNLCF - • DOE sites

• US Universities

• Industry

• International

Backbone Band width parity

Backbone band width parity

• Bulk data

Nuclear Physics (RHIC)

- • DOE sites

• US Universities

• International

12 Gbps 70 Gbps • Bulk data • Guaranteed bandwidth

• PKI / Grid

Spallation Neutron Source

High

(24x7 operation)

• DOE sites 640 Mbps

2 Gbps • Bulk data

Science Network Requirements Aggregation Summary(slide courtesy of ESNet)

CI Components

Network

Performance Infrastructure / Tools

MiddlewareControl Plane

….

Bulk Transport

2-Way Interactive

Video

Real-Time Communications

Applications

Applications call on Network Cyberinfrastructure

….

…. ….Phoebus

Netw

ork C

yberinfrastructure

Measurement Nodes

Control Plane Nodes

Internet2 Network CI Software

• Dynamic Circuit Control Infrastructure• DRAGON (with ISI, MAX)• Oscars (with ESnet)

• Middleware (Federated trust Infrastructure)• Shibboleth• Signet• Grouper• Comanage

• Performance Monitoring Infrastructure• perfSONAR (with ESnet, GEANT2 JRA1, RNP, many

others)• BWCTL, NDT, OWAMP, Thrulay

• Distributed System Infrastructure• Topology Service (with University of Delaware)• Distributed Lookup Service (with University of

Delaware, PSNC)

Internet2 Network CI Standardization

• Dynamic Circuit Control Protocol (IDC)• DICE-Control, GLIF

• Measurement Schema / Protocol• OGF NMWG• IETF IPPM• perfSONAR Consortium

• Middleware Arena• Liberty Alliance• OASIS• Possible emerging corporate consortium

• Topology Schema / Protocol• OGF NML-WG• perfSONAR Consortium• DICE-Control

Internet2’s CI Vision

• Internet2’s CI vision:• Be a networking cyber-service provider

• Be a trust cyber-service provider

• Be a CI technology developer.

Internet2’s CI Position

• Internet2’s position:• Backbone network provider• Federated trust infrastructure provider

• Forum for collaboration by members of the R&E community

• Gives Internet2 a unique vision and strategy for Cyberinfrastructure.

Internet2’s CI Definition

• Components• Supercomputing / Cycles / Computational

• Supercomputing / Storage (Non-volatile)

• Analysis / Visualization• Interconnecting Networks (Campuses, Regionals, Backbones)

• Network Cyberinfrastructure Software

Internet2’s CI Audience

• Application Software• Instrumentation / Remote Instruments / Sensors

• Data Sets

Internet2’s CI Constituencies

• Collaborators• University Members• Regional Networks• Regional CI Organizations• High Performance Computing Centers • Federal Partners• International Partners• CI Integrators

Early Thoughts: Internet2’s CI Strategy (1)

• Requirements• Informed by our membership • Agenda set by our governance mechanisms

• Offer, and in some cases develop, services and technology that are key components of a coherent CI software suite.

• For CI to work, it has to be a workable end-to-end system; Internet2 is emphasizing a systems approach towards CI.• Internet2 is offering new services such as the

Internet2 Network, InCommon, and the VO Service Center.

• Internet2 is developing and offering new technologies such as GridShib and perfSONAR.

• Internet2 may do systems integration work assembling open source communication tools into a common veneer.

Early Thoughts: Internet2’s CI Strategy (2)

• Play the role of community CI coordinator, convening community conversations.• Partner with other community coordinators (e.g. Teragrid, EDUCAUSE).

• Play a convening function in order to facilitate the development, use, and dissemination of CI (e.g. Bridging the Gap workshop).

• Take a lead in international outreach efforts at several different layers of CI.

• Work with campuses to build valuable CI.• Facilitate conversations among various federal agencies (e.g. DOE, NSF, NIH), each of which is developing its own CI, and present a consistent vision back to the campuses.

Internet2’s CI Tactics

• Target campus, national, and international audiences• Integrate campus CI into regional national/international CI

• Target Application-community CI (quasi-national)• Enable effective use of authorized resources, regardless of

where they exist• Enable integration of new resources as they become available• Facilitate interoperability of multiple, autonomous CI

providers• Take a “toolkit” approach

• Make sure it still looks like a wall jack to end user• Push for best practices for campuses

• What to do• How to do it• Community learns as a whole / avoid reinventing the wheel

• Contribute to the support structure for use of CI• Open source CI software• Centers of Excellence for various kinds of things• Training)

Questions?

• Eric Boyd• eboyd@internet2.edu