The 4th ChinaGrid Annual Conference

PROGRAMMEPROGRAMMEYantai, China, Aug 21-22, 2009

Organized by:

ChinaGrid Working Group Shandong University, China Huazhong University of

Science & Technology, China

With the sponsorship of:

Intel 浪潮集团联想（北京）有限公司

Published by:

Preface

Welcome to the 4th ChinaGrid Annual Conference (ChinaGrid 2009) which will be held in Yantai, Shandong, China, from August 21 to 22, 2009.

Grid computing presents a new trend to distributed computing for coordinating large-scale heterogeneous resources sharing and problem solving in dynamic, multi-institutional virtual organizations. China Education and Research Grid Project (ChinaGrid), under the support of Chinese Ministry of Education (MoE), aims at sharing and integrating heterogeneous mass resources distributed in Chinese universities and eventually forming a public research and education platform in China. To stimulate exchanging experiences and innovating ideas between experts and users involved in ChinaGrid community, ChinaGrid organizes the annual conference on grid computing and related areas. ChinaGrid Annual Conference provides a high profile and leading edge forum for researchers and engineers to present their latest research results on ChinaGrid.

This year we received 105 paper submissions from ChinaGrid members and other universities. All the papers were peer reviewed by at least two referees from the conference technical program committee and their colleagues. In order to allocate as many papers as possible and also ensure the overall quality of the conference, total 39 papers are eventually accepted. We believe that all these papers not only provide novel ideas, new results, work in progress and state-of-the-art techniques in this field, but also stimulate the future research activities in the area of grid computing. The conference is a result of many people’s hard work, including external reviewers, program and technical committee members. We would like to express our sincere thanks to everyone involved. In addition, the ultimate success of the conference will be judged by how well the participants learn, interact, and communicate with each other. The committees and organizers of the conference have provided a venue and created a pleasure environment to make sure these objectives to be achieved. It is now up to all of us to ensure that the conference is an outstanding success.

We wish a successful, stimulating and rewarding conference and look forward to seeing you again at future ChinaGrid Annual conferences.

   Hai Jin，Xiangxu Meng

ChinaGrid 2009 General Co-Chairs

PROGRAMME

Venue：Crowne Plaza Yantai Sea View （烟台南山皇冠假日酒店）

NO.299 GangCheng East Avenue, LaiShan District, Yantai    

Aug 20, Thuesday

8:00~23:00 Registration（酒店大堂）

Aug 21, Friday

7:30~8:30 Registration（酒店大堂）

8:30~8:50 Opening Ceremony（Aquamarine（海宝兰）Hall）

8:50~9:50Keynotes1: Market-Oriented Cloud Computing: Vision, Hype, and Reality of Delivering Computing as the 5th UtilitySpeaker: Dr. Rajkumar Buyya, The University of Melbourne and Manjrasoft, Melbourne, Australia

9:50~10:35Keynotes2: Virtualization Technology for Computing SystemSpeaker: Dr. Hai Jin, Huazhong University of Science & Technology, China

10:35~11:05Keynotes3: 英特尔领先架构加速高性能计算Speaker: Dr. Richard Dracott, Intel

11:05~11:20 Coffee break

11:20-12:40 Room1 (Aquamarine（海宝兰）Hall) Room2(Topaz（黄玉）Hall)

1

Session1:Grid Resource Management (1) (Chair:)

Session3:Data Grid(Chair:)

Reliability Analysis Approach of Grid Monitoring

Architecture

Jun Gu, Junzhou Luo

Distributed Metadata Management based on

Hierarchical Bloom Filters in Data Grid

Shihua Chen, Xiaomeng Huang, Pengzhi Xu, Weimin Zheng

FCCS: File Classification Caching Service based on

RAM Grid

ZhongAn Lao，Nong Xiao，ChangSheng Fu

STBucket: A Self-Tuning Bucket Index in DAS

Paradigm

Haocong Wang, Xiaoyong Du, Jieping Wang, Pingping Yang

To improve throughput via multi-pathing and

Parallel TCP on each path

Jinyu Zhang, Yongzhe Gui, Cheng Liu, Xiaoming Li

A Data-Intensive Workflow Scheduling Algorithm

for Grid Computing

Meng Xu, Lizhen Cui, Haiyang Wang, Yanbing Bi, Ji Bian

GridDEV: A Platform for Service Grid Evaluation

Hongbo Tian, Xiaoshe Dong, Siyuan Ma, Bingyi Li, Feifei Liu

Dynamic Replica selection services based on state

evaluation strategy

Chang ze Wu, Kai gui Wu, Ming Chen, Chunxiao YE

12:40~13:40 Lunch（一楼西餐厅）

14:00-16:00 Room1 (Aquamarine（海宝兰）Hall) Room2(Topaz（黄玉）Hall) Room3 (Pearl（珍珠）Hall)

Intel Track：Constructing Intelligent

and Efficient HPC Environments

Session1:Grid Resource Management (2) (Chair:)

Session5: Grid Applications(Chair:)

Intel Innovation on

Powering HPC Development

Dr. David Scott

A Resource Allocation

Method for Computational Grids Based on On-

line Reverse Auction

Chao Zhang, Wenyu Qu, and Zhaobin Liu, Wanlei Zhou

Realcouse2.0: From Fully

Connected Topology to Multiple Overlapping

Stars

Jinyu Zhang, Xiaoming Li

2

Constructing HPC

environment with Intel leading technology

and products

Gu Fan

Load Balancing on the

Exchanged Hypercube

Chong Yao, Keqiu Li, Kai Lin, Yanming Shen

Achievement for

Complicated Electromagnetic Environment

Simulation Application Based on ChinaGrid

Min Zhang,Haitao Jia,Shunsheng Zhang

Knowledge sharing on

benchmarking and evaluating HPC systems

Dr. Wanqing He

CampusWare: An Easy-

To-Use, Efficient and Portable Grid

Middleware for Compute-intensive

Applications

Dong Wang, Jinlei Jiang, Yongwei Wu, Guangwen Yang

Designing a Chemical Script

Editor in Grid environment with DLTK

Yongying He, Ruisheng Zhang, Ruipeng Wei, Caihua Hu, Lian Li

Double Redundant Fault-

Tolerance Service Routing Model in ESB

Bin Wu, Shijun Liu, Wei Cui

A Test Platform for

Household Appliances Based on Web Services

Zhongwen Guo, Pengpeng Chen, Keyong Hu, Yuan Feng, Feng Hong

Adaptively Construct

Banking Process with Tags Upon Services-

oriented Grid

Li Qi, Song Wu

16:00-16:15 Coffee break

16:15-18:15 Room1 (Aquamarine（海宝兰）Hall) Room2(Topaz（黄玉）Hall) Room3 (Pearl（珍珠）Hall)

Intel Track：Constructing Intelligent

and Efficient HPC Environments

Session2:Grid Job Scheduling (Chair:)

Session4&8:Grid Security & Peer-to-Peer Computing (Chair:)

Intel China Higher Multi-objective Identity-Based Sequential

3

Education Programs for ChinaGrid

Joey Guo

Optimization Approaches Using a CE-ACO

inspired strategy to improve Grid jobs

scheduling

Yi Hu, Bin Gong

Aggregate Signature from RSA

Bennian Dou, Hong Zhang, Chungen Xu, Mu Han

Inspur HPC Cluster

Management Software

Wei Jian

Backward Planning: A

Simple and Efficient Method to Improve the

Performance of List Scheduling Algorithms

Rong Deng, Changjun Jiang, Qinma Kang, Fei

Yin

On the Security of

Certificateless Authenticated Key Agreement

Protocol (CL-AK) for Grid Computing

Mengbo Hou and Qiuliang Xu

MPICH-G-DM: An

Enhanced MPICH-G with Supporting Dynamic

Job Migration

Xiaohui Wei, Hongliang Li, Dexiong Li

A Collaborative Filtering

Recommendation Model Using Polynomial

Regression Approach

Houkun Zhu, Yuan Luo, Chuliang Weng, Minglu Li

Ant Algorithm with

Execution Quality Based Prediction in Grid

Scheduling

Yiqun Zhu, Minglu Li, Chuliang Weng

An Heuristic Method for

Web-Service Program Security Testing

Gang Zhao, Weimin Zheng, Jinjing Zhao, Hua Chen

The Priority Tasks

Scheduling Algorithm Based on Grid Resource

Prediction

Hongwei Liu, Jiong Yu, Guozhong Tian, Hongcui

Gong

Event-Driven Scheduling for

P2P VoD

Bin Cheng, Hai Jin, Xiaofei Liao, Bo Li

Research on multi-QoS

on-line scheduling based on fuzzy theory in the

grid

Jing Xu, Shoubao Yang, Bin Wu, Lili Shi

Toward Data Revelation and

Scheduling Policy in Mesh-based Peer-to-Peer

Streaming

Qingchao Cai, Xuejie Zhang

4

18:30-20:00 Banquet (绿宝石宴会厅)

Aug 22, Saturday

8:30~10:35 Keynotes（绿宝石 2）

8:30~9:10Keynotes4: Evolvement from physical grid to virtual gridSpeaker: Dr. Lian Li, Lanzhou University , China

9:10~9:30Keynotes5: CPU-GPU协同加速高性能计算Speaker: Kaiyong Zhao, Inspur Group

9:30~9:50Keynotes6: 高性能计算中的异构技术Speaker: Dr. Xiaodong Shi ,Lenovo Group

9:50-10:05 Coffee break

10:05-12:05 Room1 (绿宝石 2) Room2(Topaz（黄玉）Hall)

Session6&9:Semantic Grid and Semantic Web & Cloud Computing(Chair:)

Session7:Cluster Computing and Virtual Machine(Chair:)

KeyOnto: A Hybrid Knowledge Retrieval Model in

Law Semantic Web

Biao Fan, Guangqiang Liu, Tao Liu, He Hu and Xiaoyong Du

Optimize the Performance of Virtual Machine

Checkpointing via Memory Exclusion

Haikun Liu, Hai Jin, Xiaofei Liao

Ontology Based Semantic Relation Verification for A Reliable Parallel Interval Global Optimization

5

TCM Semantic Grid

Xiaogang Zhang,Huajun Chen, Jun Ma, Jinhuo TaoAlgorithm Based On Mind Evolutionary Computation

Yongmei Lei,Shaojun Chen

Research of Ontology Modeling in Structure

Engineering Grid

LONG hao, LU Hai, Di Rui-Hua

Correctness Analysis based on Testing and

Checking for OpenMP Programs

Jianjiang Li, Dan Hei, Lin Yan

MReC4.5: C4.5 ensemble classification with

MapReduce

Gongqing Wu, Haiguang Li, Xuegang Hu, Yuanjun Bi, Jing Zhang, Xindong Wu

Research on The Performance of xVM Virtual

Machine Based on HPCC

Tiezhu Zhao, Yilong Ding, Verdi March, Shoubin Dong, Simon See

AS-M:Resource Allocation Strategy based on

Market Mechanism in Cloud Computing

Xindong YOU, Xianghua XU, Jian WAN, Dongjin YU

The Application of Virtual Machine on System

Security

Siqin Zhao,Kang Chen,Weimin Zheng

12:05-13:30 Closing Ceremony & Lunch（一楼西餐厅）

6

Table of Contents

KEYNOTE SPEECHES..........................................................................................................................1

MARKET-ORIENTED CLOUD COMPUTING: VISION, HYPE, AND REALITY OF DELIVERING COMPUTING

AS THE 5TH UTILITY..............................................................................................................................1

EVOLVEMENT FROM PHYSICAL GRID TO VIRTUAL GRID.........................................................................3

VIRTUALIZATION TECHNOLOGY FOR COMPUTING SYSTEM...................................................................4

英特尔领先架构加速高性能计算.......................................................................................................5

CPU-GPU 协同加速高性能计算.........................................................................................................5

高性能计算中的异构技术...................................................................................................................5

SESSION 1: GRID RESOURCE MANAGEMENT............................................................................6

RELIABILITY ANALYSIS APPROACH OF GRID MONITORING ARCHITECTURE........................................6

FCCS: FILE CLASSIFICATION CACHING SERVICE BASED ON RAM GRID.............................................6

TO IMPROVE THROUGHPUT VIA MULTI-PATHING AND PARALLEL TCP ON EACH PATH........................7

LOAD BALANCING ON THE EXCHANGED HYPERCUBE...........................................................................7

A RESOURCE ALLOCATION METHOD FOR COMPUTATIONAL GRIDS BASED ON ON-LINE REVERSE

AUCTION................................................................................................................................................8

DOUBLE REDUNDANT FAULT-TOLERANCE SERVICE ROUTING MODEL IN ESB...................................8

CAMPUSWARE: AN EASY-TO-USE, EFFICIENT AND PORTABLE GRID MIDDLEWARE FOR COMPUTE-

INTENSIVE APPLICATIONS......................................................................................................................9

GRIDDEV: A PLATFORM FOR SERVICE GRID EVALUATION.................................................................9

SESSION 2: GRID JOB SCHEDULING............................................................................................10

MULTI-OBJECTIVE OPTIMIZATION APPROACHES USING A CE-ACO INSPIRED STRATEGY TO IMPROVE

GRID JOBS SCHEDULING.......................................................................................................................10

BACKWARD PLANNING: A SIMPLE AND EFFICIENT METHOD TO IMPROVE THE PERFORMANCE OF

LIST SCHEDULING ALGORITHMS..........................................................................................................10

MPICH-G-DM: AN ENHANCED MPICH-G WITH SUPPORTING DYNAMIC JOB MIGRATION..............11

ANT ALGORITHM WITH EXECUTION QUALITY BASED PREDICTION IN GRID SCHEDULING................11

RESEARCH ON MULTI-QOS ON-LINE SCHEDULING BASED ON FUZZY THEORY IN THE GRID................12

THE PRIORITY TASKS SCHEDULING ALGORITHM BASED ON GRID RESOURCE PREDICTION..............12

SESSION 3: DATA GRID.....................................................................................................................13

DISTRIBUTED METADATA MANAGEMENT BASED ON HIERARCHICAL BLOOM FILTERS IN DATA GRID

.............................................................................................................................................................13

STBUCKET: A SELF-TUNING BUCKET INDEX IN DAS PARADIGM......................................................13

A DATA-INTENSIVE WORKFLOW SCHEDULING ALGORITHM FOR GRID COMPUTING.........................14

DYNAMIC REPLICA SELECTION SERVICES BASED ON STATE EVALUATION STRATEGY........................14

SESSION 4: GRID SECURITY...........................................................................................................15

IDENTITY-BASED SEQUENTIAL AGGREGATE SIGNATURE FROM RSA.................................................15

ON THE SECURITY OF CERTIFICATELESS AUTHENTICATED KEY AGREEMENT PROTOCOL (CL-AK)

FOR GRID COMPUTING.........................................................................................................................15

A COLLABORATIVE FILTERING RECOMMENDATION MODEL USING POLYNOMIAL REGRESSION

I

APPROACH............................................................................................................................................16

AN HEURISTIC METHOD FOR WEB-SERVICE PROGRAM SECURITY TESTING......................................16

SESSION 5: GRID APPLICATIONS..................................................................................................17

A TEST PLATFORM FOR HOUSEHOLD APPLIANCES BASED ON WEB SERVICES..................................17

ADAPTIVELY CONSTRUCT BANKING PROCESS WITH TAGS UPON SERVICES-ORIENTED GRID...........17

ACHIEVEMENT FOR COMPLICATED ELECTROMAGNETIC ENVIRONMENT SIMULATION APPLICATION

BASED ON CHINAGRID.........................................................................................................................18

DESIGNING A CHEMICAL SCRIPT EDITOR IN GRID ENVIRONMENT WITH DLTK.................................18

REALCOUSE2.0: FROM FULLY CONNECTED TOPOLOGY TO MULTIPLE OVERLAPPING STARS............19

SESSION 6: SEMANTIC GRID AND SEMANTIC WEB................................................................20

KEYONTO: A HYBRID KNOWLEDGE RETRIEVAL MODEL IN LAW SEMANTIC WEB............................20

ONTOLOGY BASED SEMANTIC RELATION VERIFICATION FOR TCM SEMANTIC GRID.......................20

RESEARCH OF ONTOLOGY MODELING IN STRUCTURE ENGINEERING GRID........................................21

SESSION 7: CLUSTER COMPUTING AND VIRTUAL MACHINE.............................................22

OPTIMIZE THE PERFORMANCE OF VIRTUAL MACHINE CHECKPOINTING VIA MEMORY EXCLUSION. .22

A RELIABLE PARALLEL INTERVAL GLOBAL OPTIMIZATION ALGORITHM BASED ON MIND

EVOLUTIONARY COMPUTATION...........................................................................................................22

CORRECTNESS ANALYSIS BASED ON TESTING AND CHECKING FOR OPENMP PROGRAMS.................23

RESEARCH ON THE PERFORMANCE OF XVM VIRTUAL MACHINE BASED ON HPCC.........................23

THE APPLICATION OF VIRTUAL MACHINE ON SYSTEM SECURITY......................................................24

SESSION 8: PEER-TO-PEER COMPUTING...................................................................................25

EVENT-DRIVEN SCHEDULING FOR P2P VOD.......................................................................................25

TOWARDS DATA REVELATION AND SCHEDULING POLICY IN MESH-BASED PEER-TO-PEER

STREAMING..........................................................................................................................................25

SESSION 9: CLOUD COMPUTING..................................................................................................26

MREC4.5: C4.5 ENSEMBLE CLASSIFICATION WITH MAPREDUCE.......................................................26

RAS-M:Resource Allocation Strategy based on Market Mechanism in Cloud Computing...............26

II

Keynote Speeches

Market-Oriented Cloud Computing: Vision, Hype, and Reality of

Delivering Computing as the 5th Utility

Rajkumar BuyyaThe University of Melbourne and Manjrasoft, Melbourne, Australia

Abstract

Computing is being transformed to a model consisting of services that are commoditised and delivered in a manner similar to utilities such as water, electricity, gas, and telephony. In such a model, users access services based on their requirements without regard to where the services are hosted. Several computing paradigms have promised to deliver this utility computing vision and they include Grid computing, P2P computing, and more recently Cloud computing. The latter term denotes the infrastructure as a “Cloud” in which businesses and users are able to access applications from anywhere in the world on demand. Cloud computing delivers infrastructure, platform, and software (application) as services, which are made available as subscription-based services in a pay-as-you-go model to consumers. These services in industry are respectively referred to as Infrastructure as a Service (Iaas), Platform as a Service (PaaS), and Software as a Service (SaaS). To realize Cloud computing, vendors such as Amazon, HP, IBM, and Sun are starting to create and deploy Clouds in various locations around the world. In addition, companies with global operations require faster response time, and thus save time by distributing workload requests to multiple Clouds in various locations at the same time. This creates the need for establishing a computing atmosphere for dynamically interconnecting and provisioning Clouds from multiple domains within and across enterprises. There are many challenges involved in creating such Clouds and Cloud interconnections.

This keynote talk (1) presents the 21st century vision of computing and identifies various IT paradigms promising to deliver the vision of computing utilities; (2) defines the architecture for creating market-oriented Clouds and computing atmosphere by leveraging technologies such as VMs; (3) provides thoughts on market-based resource management strategies that encompass both customer-driven service management and computational risk management to sustain SLA-oriented resource allocation; (4) presents the work carried out as part of our new Cloud Computing initiative, called Cloudbus: (i) Aneka, a software system for providing PaaS within private or public Clouds and supporting market-oriented resource management, (ii) internetworking of Clouds for dynamic creation of federated computing environments for scaling of elastic applications, (iii) creation of 3rd party Cloud brokering services for content delivery network and e-Science applications and their deployment on capabilities of IaaS providers such as Amazon and Nirvanix

1

along with Grid mashups, and (iv) CloudSim supporting modelling and simulation of Clouds for performance studies; and (5) concludes with the need for convergence of competing IT paradigms for delivering our 21st century vision along with pathways for future research.

Dr. Rajkumar Buyya is an Associate Professor of Computer Science and

Software Engineering; and Director of the Cloud Computing and

Distributed Systems (CLOUDS) Laboratory at the University of

Melbourne, Australia. He is also serving as the founding CEO of

Manjrasoft Pty Ltd., a spin-off company of the University,

commercialising its innovations in Grid and Cloud Computing. He has

authored over 280 publications and three books. The books on emerging

topics that Dr. Buyya edited include, High Performance Cluster

Computing (Prentice Hall, USA, 1999), Content Delivery Networks (Springer, 2008) and Market-

Oriented Grid and Utility Computing (Wiley, 2009). Dr. Buyya has contributed to the creation of

high-performance computing and communication system software for Indian PARAM

supercomputers. He has pioneered Economic Paradigm for Service-Oriented Distributed

Computing and developed key Grid and Cloud Computing technologies such as Gridbus and

Aneka that power the emerging e-Science and e-Business applications. In this area, he has

published hundreds of high quality and high impact research papers that are well referenced. The

Journal of Information and Software Technology in Jan 2007 issue, based on an analysis of ISI

citations, ranked Dr. Buyya's work (published in Software: Practice and Experience Journal in

2002) as one among the "Top 20 cited Software Engineering Articles in 1986-2005". He received

the Chris Wallace Award for Outstanding Research Contribution 2008 from the Computing

Research and Education Association of Australasia. He is the recipient of 2009 IEEE Medal for

Excellence in Scalable Computing.

2

Evolvement from physical grid to virtual grid

Lian Li

College of Computer and information engineering

Hefei University of Technology, China

Abstract

Grid is rapidly grown as a common computing platform within lasted twenty years, which integrals various discrete resources into a uniform system to be shared by numerical users, not only the utilizers but also developers, without perception of allochthonous and heterogeneity. Up to date, the grid is evolving from physics grid to virtual grid, i.e. from a grid programming oriented the resources to a grid programming oriented services. It results the application of grid more fields and scopes. The virtual technique has promoted greatly the emerging of many standards on grid, and catalyzes a new programming for grid application, which is called orchestration of services, The virtual grid change the fashion and concept of using computers. In this report, we discuss some theoretical points of virtual technique, the current situation of grid standards, and the problems in grid system developing.

Prof. Li Lian is faculty in school of information science and engineering,

Lanzhou University, Director of Open-source software and real-time system

engineering researching centre of Education Ministry (OSSRT), and

computer mathematic and web computing Laboratory at Lanzhou University,

China. He has authored over 60 publications. The recent researching include,

virtualization technique, grid computing, pervasive computing, network

algorithms, algorithms of mathematic problem, mathematical theory on

computer science. Prof. Li Lian has developing an e-science system on chemistry application

supported by National Natural Science Foundation of China(NNSF). He is associate director of

theoretical computer science branch, China Association of Computer Science, and director of

steering committee for computer curricula on science and technique, Education Ministry of China.

3

Virtualization Technology for Computing System

Hai JinSchool of Computer Science and Technology

Huazhong University of Science and Technology, China

Abstract

Virtualization technology (VT) being a re-emerging technology has become a hot topic in recent years. It is a decoupling technique that separates system software from hardware platform while making applications run pervasively. Many academic institutes and research labs from industries have devoted great research efforts in various aspects. In 2007, Chinese Ministry of Science and Technology initialized a basic research project (973 project) with 6 universities and 2 research institutes on virtualization technology for computing system in many topics, such as architecture design philosophy of virtualization technology, VT design for single computing system, VT design for multiple computing systems, user environment for VT, security, reliability and trust issue related to VT, and performance evaluation and benchmarks for computing system with VT. In this talk, we will give insight for this project in details, including its motivation, goal, and research issue etc. Some preliminary research results with this project will also be illustrated.

Dr. Hai Jin is a professor of computer science and engineering at the

Huazhong University of Science and Technology (HUST) in China. He

is now Dean of the School of Computer Science and Technology at

HUST. Jin received his PhD in computer engineering from HUST in

1994. In 1996, he was awarded a German Academic Exchange Service

fellowship to visit the Technical University of Chemnitz in Germany. Jin

worked at The University of Hong Kong between 1998 and 2000, and as

a visiting scholar at the University of Southern California between 1999

and 2000. He was awarded Excellent Youth Award from the National

Science Foundation of China in 2001. Jin is the chief scientist of

ChinaGrid, the largest grid computing project in China.

Jin is a senior member of the IEEE and a member of the ACM. Jin is the member of Grid

Forum Steering Group (GFSG). He has co-authored 15 books and published over 400 research

papers. His research interests include computer architecture, virtualization technology, cluster

computing and grid computing, peer-to-peer computing, network storage, and network security.

Jin is the steering committee chair of International Conference on Grid and Pervasive Computing

(GPC), Asia-Pacific Services Computing Conference (APSCC). Jin is a member of the steering

committee of the IEEE/ACM International Symposium on Cluster Computing and the Grid

(CCGrid), the IFIP International Conference on Network and Parallel Computing (NPC), and the

International Conference on Grid and Cooperative Computing (GCC), International Conference

on Autonomic and Trusted Computing (ATC), International Conference on Ubiquitous

4

Intelligence and Computing (UIC).

英特尔领先架构加速高性能计算

Richard DracottGeneral Manager – High Performance Computing (HPC)

Abstract

英特尔平台为高性能计算的行业用户提供智能的性能及从各个方面为高性能

计算生态系统的发展提供强大的支持。这个演讲会谈到英特尔在高性能计算市场

中的发展， 从产品角度会回顾全新 Nehalem架构的至强 5500带给高性能计算

市场的影响，明年初基于 Nehalem架构的高可靠，超大内存的多路 Nehalem-EX

平台对市场的影响；从市场角度会谈到英特尔在全球 Volume 高性能市场及高

端超算市场的所开展的不同的工作；从软件的角度会谈到高性能计算应以应用

为本，如何利用英特尔软件工具及行业专家资源来充分发挥出计算平台的潜能；

从未来发展趋势的角度会谈到如何满足未来不断增长的计算需求，应对能耗及

应用并行化的挑战。

Richard is the General Manager of High Performance Computing in the Server Platforms

Group.  He is responsible for driving a cross organizational team to grow Intel’s overall business

in this fast growing market segment including silicon, platforms and software products, in addition

to industry initiatives and collaboration.

Richard joined Intel in Canada in 1982 and has held a variety of international sales, marketing

and strategy positions in Europe and North America, including Director of Marketing for the

Microprocessor Products Group, Director of Marketing for the Desktop Product Group and

Business Manager for Intel Online Services Europe. From 2001 to 2004, he was the General

Manager of Marketing & Planning for the Enterprise Platforms Group, focused on the Intel®

Itanium® 2 processor and Intel® Xeon(tm) processor families for servers and workstations.  Until

November 2006, he was the Director of Strategic Marketing for the End-User Platform Integration

organization, focused on engaging end users to help drive future product planning.

Prior to Intel, Richard was in software development for aerospace and communications, and

holds a BSC in Mathematics from the University of Southampton in England.

5

CPU-GPU协同加速高性能计算

Kaiyong ZhaoInspur Group

高性能计算中的异构技术

Dr. Xiaodong ShiLenovo Group

6

Session 1: Grid Resource Management

Reliability Analysis Approach of Grid Monitoring ArchitectureJun Gu1,2, Junzhou Luo1

1School of Computer Science and EngineeringSoutheast University, Nanjing, China

2School of Computer Science and TechnologyChina University of Mining and Technology, Xuzhou, China

AbstractGrid monitoring, as an important part of any grid systems, is needed to query the

state of grid resources and match user requirements with available grid resources. In order to ensure the availability of grid monitoring, the reliability imposed by software or hardware failure happened with unpredictable probability must be assessed. This paper contributes to study the reliability analysis approach of grid monitoring in the context of grid monitoring architecture (GMA) that has been de facto standards for many areas of grid computing. Failure types and contributing factors in GMA are analyzed, which are likely to take place in comprised components, channels or process behaviors. Then, the respective evaluation equations are suggested via Markov procedure, queue model, and probability theory. Furthermore, the reliability issue of hierarchical GMA is discussed based on four basic architectural relations. Numerical examples are given to illustrate the proposed computing equations. The results show that our approach is feasible.

FCCS: File Classification Caching Service based on RAM GridZhongAn Lao, Nong Xiao, ChangSheng Fu

School of ComputerNational University of Defense Technology

Changsha, ChinaAbstract

Memory intensive applications and I/O intensive applications often suffer from the poor performance of disk swapping when memory is inadequate.RAM Grid, which combines Network Memory, Service-oriented Computing and Grid Computing technology, focuses on solving these problems. Being excellent at providing shared memory to improve system performance, RAM Grid can’t guarantee busy node load balance or optimal file remote cache efficiency. After studying data placement policy of large-scale network storage systems, we propose a file classification caching service (FCCS) based on RAM Grid, which promises fairness and high availability, to give a method that tries to solve those shortages. As experiment results show: FCCS improve system performance greatly.

7

To improve throughput via multi-pathing and Parallel TCP on each pathJinyu Zhang, Yongzhe Gui, Cheng Liu, Xiaoming Li

School of Electronics Engineering and Computer SciencePeking University, China

AbstractParallel TCP, which opens multiple TCP con-nections over a single direct path,

and Multi-Pathing, whichconcurrently uses multiple disjointed paths to transfer data, have both been proved to be effective methods to improve end-to-end throughput. How much throughput can we ultimately achieve between a source and a destination if we use multiple overlay paths and open multiple TCP connections on each used path?

In order to find all possible overlay paths of good quality between a source and a destination, a path probing process similar to the path discovery protocol of IEEE 802.5 is started by the destination. A probing packet(a TCP connection request followed by padding data) is flooded across an overlay between the destination and the source. Intermediate overlay nodes selectively accept and forward probing packets. If a probing pack is accepted, a corresponding TCP connection is created. Trade-offs then are made between reducing the probing traffic and keeping multiple TCP connections on each path. The source strips data into small packets and adaptively assigns them to selected overlay paths according to the changing quality of each path.

This proposed data transfer technology is evaluated within an overlay that consists of 15 servers on the Internet in China, across 3 different autonomous systems. Experiments show that with this technology, 54% of the measured samples yield a throughput larger than 60Mb/s, which is 60% of the band-width that could be possibly obtained(the access bandwidth is 100Mb/s for all servers). Comparing with direct path and Parallel TCP, only less than 1% and 25% of the measured samples reach the same level of throughput respectively.

Load Balancing on the Exchanged HypercubeChong Yao, Keqiu Li, Kai Lin, Yanming Shen

Department of Computer Science and EngineeringDalian University of Technology, Dalian, China

AbstractThe Exchanged Hypercube is an interconnection network which is obtained by

systemically removing some links from a binary hypercube. In parallel systems, load balancing is a very important factor which can affect the performance of the whole system. So distributing tasks evenly on processors is essential for multiprocessor computing systems. Based on the classical DE-based algorithm, in this paper we propose a load balancing algorithm for the Exchanged Hypercube architecture. We also theoretically prove the correctness of the proposed algorithm. Finally, we use a case study to further explain our algorithm.

8

A Resource Allocation Method for Computational Grids Based on On-line Reverse Auction

Chao Zhang1, Wenyu Qu1, Zhaobin Liu1, Wanlei Zhou2

1School of Information Science and TechnologyDalian Maritime University, Dalian, China

2School of Computing and Mathematics, Deakin University, AustraliaAbstract

Resource allocation and task scheduling are two key technologies in grid computing system. The market-based resource allocation model is considered as a good one. In this paper, an on-line reverse auction method of resource allocation for computational grids was proposed to solve the problem of resource management considering the dynamic characteristics of computing resources in the computational grid environment and the advantages of economics mechanism. In this method, the current price can be set using former bids. And bidders arriving one by one the on-line buyer must be required to make a decision immediately about each bid as it is received. Then we prove that the algorithm is incentive compatible and simulate the auction protocol in Gridsim to evaluate its communication demand.

Double Redundant Fault-Tolerance Service Routing Model in ESBBin Wu1, Shijun Liu1, Wei Cui2

1School of Computer Science and TechnologyShandong University, Jinan, China

2 Shandong Jianzhu University, Jinan, ChinaAbstract

With the development of the Service Oriented Architecture (SOA), the Enterprise Service Bus (ESB) is becoming more and more important in the management of mass services. The main function of it is service routing which focuses on delivery of message among different services. At present, some routing patterns have been implemented to finish the messaging, but they are all static configuration service routing. Once one service fails in its operation, the whole service system will not be able to detect such fault, so the whole business function will also fail finally.

In order to solve this problem, we present a double redundant fault tolerant service routing model. This model has its own double redundant fault tolerant mechanism and algorithm to guarantee that if the original service fails, another replica service that has the same function will return the response message instead automatically. The service requester will receive the response message transparently without taking care where it comes from. Besides, the state of failed service will be recorded for service management.

At the end of this article, we evaluated the performance of double redundant fault tolerant service routing model. Our analysis shows that, by importing double redundant fault tolerance, we can improve the fault-tolerant capability of the services

9

routing apparently. It will solve the limitation of existent static service routing and ensure the reliability of messaging in SOA.

10

CampusWare: An Easy-To-Use, Efficient and Portable Grid Middleware for Compute-intensive Applications

Dong Wang, Jinlei Jiang, Yongwei Wu, Guangwen YangDepartment of Computer Science and Technology

Tsinghua National Laboratory for Information Science and TechnologyTsinghua University, Beijing, China

Abstract

This paper describes the design and implementation of CampusWare, a lightweight grid middleware for the compute-intensive applications in campus environment. Though there are many grid middlewares available, they are unnecessarily complex for use in the campus environment where the main applications are compute-intensive, and the main requirements are the convenience and efficiency in clusters management and job submissions. To deal with the problem, CampusWare proposes a “fast job” concept and provides a two-layer middleware architecture as well as a three-layer user account hierarchy. Compared with the existing grid middlewares, its deployment, configuration and usage are all simplified, which makes it easy-to-use, efficient and portable.

GridDEV: A Platform for Service Grid Evaluation

Hongbo Tian, Xiaoshe Dong, Siyuan Ma, Bingyi Li, Feifei LiuDepartment of Computer Science and Technology

Xi'an Jiaotong University, Xi'an, China

Abstract

GridDEV is a DEVS-based platform for evaluation of service grid. Through the analysis of SOA, we characterize important features of service grid and select four components as the base of grid modeling. With DSDEVS, this paper models the basic components selected, as well as the relationships between them. As a result, the model fully formalizes the gird system; hence facilitates the design and optimization of grid services. Based on this model, we map the DEVS component into elements of simulator development library SimGrid, and construct the platform GridDEV for grid performance and reliability evaluation.

11

Session 2: Grid Job Scheduling

Multi-objective Optimization Approaches Using a CE-ACO inspired strategy to improve Grid jobs scheduling

Yi Hu1, Bin Gong2

1Department of Computer Science and TechnologyShandong University, Jinan, China

2Department of Computer Science and TechnologyShandong University, Jinan, China

AbstractGrid scheduling is one of the most crucial issue in a grid environment because it

strongly affects the performance of the whole system. Taking into account that the issue of allocating jobs on resources is a combinatorial optimization problem, a NP-complete problem, several heuristics have been proposed to provide good performance. In this paper, the proposed approach considers a stochastic optimization called the cross entropy method. The CE method is used to tackle efficiently the initialization sensitiveness problem associated with ant colony algorithm for multi-objective scheduling ， which accelerates the convergence rate and improves the ability of searching an optimum solution. Simulation shows that it performs better than the ACO in the integrated performances.

Backward Planning: A Simple and Efficient Method to Improve the Performance of List Scheduling Algorithms

Rong Deng, Changjun Jiang, Qinma Kang, Fei YinDepartment of Computer Science and Engineering

Tongji University, Shanghai,ChinaAbstract

The problem of scheduling task graph which is represented by DAG (Directed Acyclic Graph) in Hetero-geneous Environment is not a new problem. However, up to now nearly all of the Heuristics use Forward Planning methods. This paper introduces novel BackWard Planning (BWP) procedure in this field. Quite different from Forward Planning methods, BWP always delays the tasks’ starttime as late as possible, in an attempt to delay the starttime of the entry nodes and shorten the makespan of the schedule in the end.

Combining BWP and traditional List Scheduling Algorithm, this paper proposes a two pass Forward/Backward (F/B) scheduling technique which can improve the performance of the original algorithm significantly. Using famous HEFT algorithm as example, we evaluate and compare the performance of two scheduling technique: our F/B technique and traditional insertion based scheduling technique used by HEFT and

12

some other existing list scheduling algorithms. The experimental results indicate that F/B technique outperforms another in all the metrics considered in our examination.

13

MPICH-G-DM: An Enhanced MPICH-G with Supporting Dynamic Job Migration

Xiaohui Wei, Hongliang Li, Dexiong LiCollege of Computer Science and Technology

Jilin University, ChangChun, ChinaAbstract

Grid is attracting more and more attentions by its massive computational capacity. Tools like Globus Toolkit and MPICH-G2 have been developed to help scientists to facilitate their researches. As a Grid-enabled implementation of MPI, MPICH-G2 helps developers to port parallel applications to cross-domain environment. Since the current computationally-intensive parallel applications, especially long-running tasks, require high availability as well as high performance computing platform, dynamic job migration in Grid environment has became an essential issue. In this study, we present a dynamic job migration enabled MPICH-G2 version, MPICH-G-DM. We use Virtual Job Model (VJM) to reserve resources for the migrating jobs in advance to improve the efficiency of the system. An Asynchronous Migration Protocol (AMP) is proposed to enable the migrating sub jobs to checkpoint/restart and update their new addresses concurrently without a global synchronization. In order to reduce the communicating overhead of job migration, MPICH-G-DM minimized the number of control messages among domains to O(N). Experiment results show that MPICH-G-DM is effective and reliable.

Ant Algorithm with Execution Quality Based Prediction in Grid Scheduling

Yiqun Zhu, Minglu Li, Chuliang WengDepartment of Computer Science,

Shanghai Jiao Tong University, Shanghai, ChinaAbstract

Task scheduling is one important job in Grid computing and also a hard and complex problem. We have quite a few algorithm of task scheduling in the past researches. Ant Algorithm is a heuristic algorithm. The inherent parallelism and scalability make this algorithm meet the requirement of complex task scheduling in Grid computing.

In this paper, we will propose an improved ant algorithm in our power grid environment. We made several improvements on the calculation of pheromone and task issue method as well. The improved algorithm becomes more sensitive with the power grid environment and more robust with heavy workload.

14

Research on multi-QoS on-line scheduling based on fuzzy theory in the grid

Jing Xu, Shoubao Yang, Bin Wu, Lili ShiDepartment of Computer Science and Technology

University of Science and Technology of China Hefei, China

AbstractIn the background of widely use of grid applications, users would no longer be

satisfied with the current grid infrastructure services of “do my best”, and hope the grid system is able to provide those services which can guarantee quality. Furthermore, as the expansion of user group, the service should be much easier to use and correspond with the applications’ particular requirements. Users’ QoS (Quality of Service) requirements are of subjective, fuzzy and incompletely considering, so aiming at these problems this paper proposes a method named FuzzyQoS. It considers users’ multi-QoS requirements using fuzzy decision-making theory based on their preferences in online scheduling. This method not only normalizes the multi-QoS requirements which have the features of subjectivity and fuzziness, but also highlights the QoS requirements of user preferences. It is shown that, with the same makespan, the degree of user satisfaction is close to other multi-QoS methods and significantly improved comparing with one-dimensional QoS scheduling method. What is the most important, this method does not require users to be sufficiently professional. It is simple, effective and sensitive to the users’ preferences.

The Priority Tasks Scheduling Algorithm Based on Grid Resource Prediction

Hongwei Liu1, Jiong Yu1,2, Guozhong Tian3,4 Hongcui Gong1

1School of Information Science and EngineeringXinjiang University, Urumqi, China

2School of Computer Science and TechnologyBeijing Institute of Technology, Beijing, China

3College of Computer Science and TechnologyBeijing University of Technology, Beijing, China

4Dept of Computer Engineering, Xinjiang Polytechnic College, Urumqi, ChinaAbstract

According to the dependence and deadline of the grid workflow tasks, the effective degrees and MIPS of the grid resources, a new algorithm called the priority tasks scheduling algorithm based on the grid resource prediction is presented. The algorithm uses DAG to find the critical path, obtain the deadline of every task and compute their PRI (priority). The algorithm takes the below problems into consideration: the request of user, the type of resources and re-scheduling of failed tasks. The result shows that the algorithm is effective.

15

Session 3: Data Grid

Distributed Metadata Management based on Hierarchical Bloom Filters in Data Grid

Shihua Chen, Xiaomeng Huang, Pengzhi Xu, Weimin ZhengDepartment of Computer Science and Technology

Tsinghua National Laboratory for Information Science and TechnologyTsinghua University Beijing, China

AbstractDistributed metadata management is an important issue in the design and

implementation of Data Grid. The key challenge lies in the strategies of metadata synchronization and the representation of the distributed metadata. We have designed a Hierarchical Bloom Filter, which consists of two level Bloom filters, to facilitate the metadata management. A Recent Bloom Filter at the top level is based on the list of recent accessed files while a Summary Bloom Filter at the bottom level represents the set of entire files. Furthermore, we propose a novel update scheme to make Recent Bloom Filters synchronized among metadata servers. Each metadata server could use the Hierarchical Bloom Filters to reduce the update frequency and the network overhead. The experimental results show that the Hierarchical Bloom Filters improve the performance and scalability of Data Grid markedly.

STBucket: A Self-Tuning Bucket Index in DAS Paradigm

Haocong Wang, Xiaoyong Du, Jieping Wang, Pingping YangKey Laboratory of Data Engineering and Knowledge Engineering, MOE

School of Information, Renmin University of China, Beijing, ChinaAbstract

In the Database-As-a-Service (DAS) paradigm, data owners outsource their data to the third-party service provider. Since the service provider is untrusted, the data should be encrypted before outsourced. Various approaches have been proposed to query on encrypted data, among which bucket based method is effective. However, previous researches just look at the data distribution with respect to a given workload, which is ineffective in changing workload behaviors. In this paper, we propose a Self-Tuning Bucket scheme: STBucket. By gathering and analyzing query feedback, STBucket achieves adaptation to workload through online bucket splitting and merging. Experimental results show that STBucket is workload aware and performs well with reasonable overhead.

16

A Data-Intensive Workflow Scheduling Algorithm for Grid Computing

Meng Xu, Lizhen Cui, Haiyang Wang, Yanbing Bi, Ji BianSchool of Computer Science and Technology

Shandong University, Jinan, China

Abstract

The data-intensive workflow in scientific and enterprise grids has gained popularity in recent times. Data-intensive workflow needs to access, process and transfer large datasets that may each be replicated on different data hosts. Because of the large data sets, the execution time is bounded by the cost of data transfer. Minimizing the time of transferring these datasets to the computational resources where the tasks of workflow are executed requires that appropriate computational and data resources be selected. In this paper, we introduce an algorithm MDTT to select the resource set which the task should be mapped. Our experiments show that our algorithm is able to minimize the total makespan of data-intensive workflow and the time of data transferring.

Dynamic Replica selection services based on state evaluation strategy

Chang ze Wu, Kai gui Wu, Ming Chen, Chunxiao YECollege of computer science

ChongQing University, ChongQing,China

Abstract

Data Replication in large scale distributed systems was introduced for enhancing data availability, reducing network flow and access latency. Efficient selecting the best replica to minimize access latency remains a fundamental challenge. Existing algorithms mostly use the previous history of replica transfers stored locally to estimate the replica access time for selection replica. It will not always give the best result especially stale access replica log. In this research, we propose a replica selection services according to evaluate replication state. To selected the best replica from replication state information according to replica node state fuzzification evaluation. The replica node state’s duration time called “state life-cycle”. It is calculated by grey prediction model. At last, the experimental results show that proposed replica selection services can improve performance to reducing access latency.

17

Session 4: Grid Security

Identity-Based Sequential Aggregate Signature from RSA

Bennian Dou, Hong Zhang, Chungen Xu, Mu HanSchool of Computer Science

Nanjing University of Science & Technology, Nanjing, China

Abstract

Sequential aggregate signature (SAS) schemes allow multiple signers to sequentially produce a short signature of different messages and also allow signers to attest to these messages as well as the order in which they signed. Identity-based signature (IBS) schemes allow a signer to sign a message, in which the signature can be verified by his identity. At CCS 2007, Boldyreva et al. proposed an identity-based sequential aggregate signature from pairings on elliptic curves, which have the merits of both SAS and IBS schemes .In this paper, we present and prove secure an identity-based sequential aggregate signature scheme, which is not based on pairings but based on RSA.

On the Security of Certificateless Authenticated Key Agreement Protocol (CL-AK) for Grid Computing

Mengbo Hou and Qiuliang XuSchool of Computer Science and Technology

Shandong University, Jinan, China

Abstract

Grid Security Infrastructure (GSI) provides an efficient mechanism to solve security problems using conventional public key infrastructure (PKI). The notion of certificateless public key cryptography gives another efficient cryptographic primitive to support Grid security services. In the recent work, Wang et al. proposed the first certificateless authentication and key agreement protocol (CL-AK) for Grid computing based on the Diffie-Hellman key agreement protocol and certificateless public key cryptography, which fits well with the GSI and provides a more lightweight key management approach for entity or data authentication and confidential protection. The authors declare that the protocol achieves many security goals. However, we found the scheme cannot withstand key compromise impersonation attack and key replicating attack, thus it doesn’t possess some desirable security attributes, such as key compromise impersonation resilience and key

18

integrity. We analyze the key replicating attack against the protocol in the BR93 security model in more detail.

19

A Collaborative Filtering Recommendation Model Using Polynomial Regression Approach

Houkun Zhu, Yuan Luo, Chuliang Weng, Minglu LiComputer Science and Engineering Department Shanghai Jiao Tong University, Shanghai, China

Abstract

In gird environment, collaborative filtering (CF) could be used for security recommendation when grid users face plenty of unknown security grid services. Also, CF recommender systems could be employed in the virtual machines managing platform to measure the creditability of each virtual machine. In this study, a polynomial regression based recommendation model on the basis of typical user-based CF is built to make security recommendation. In the model, a cluster of recommendation algorithms based on polynomial regression are derived according to various regression orders and dataset sizes. From our experiments, three significant conclusions are discovered in this model. Firstly, algorithms with lower regression orders make better predictions. Secondly, among algorithms with each fixed regression order, the best one satisfies that its dataset size is equal to its regression order in general. Thirdly, selecting appropriate regression order and dataset size could enhance recommendation quality.

An Heuristic Method for Web-Service Program Security TestingGang Zhao1,2, Weimin Zheng1, Jinjing Zhao2, Hua Chen

1Department of Computer Science and TechnologyTsinghua University, Beijing

2Beijing Institute of System Engineering, Beijing, China

Abstract

The security of the web-service program is a very significant facet in the grid computing envroinment. A fuzzer is a program that attempts to discover security vulnerabilities by sending random input to an application. How to efficiently reduce the fuzzing data scale with the assurance of high fuzzing veracity and vulnerability coverage is a very important issue for its effective practice. In this paper, aimed at the web-service program, a new heuristic method for fuzzing data generation named as H-Fuzzing is be presented, which has high program executing path coverage with the information from the static analysis and dynamic property of the program. The main thought of H-Fuzzing is collecting the information of the key branch predications and building its relations with the input variables in order to supervise the dimension reducing of the fuzzing data aggregation.

20

Session 5: Grid Applications

A Test Platform for Household Appliances Based on Web ServicesZhongwen Guo, Pengpeng Chen, Keyong Hu, Yuan Feng, Feng Hong

College of Information Science and Engineering Ocean University of China, Qingdao, China

AbstractTesting of household appliances in the manufacturing enterprises often suffers

from low efficiency. The adoption of the web services and grid technology can provide the promising future for improving the test and development efficiency of products. This paper describes a complete test platform for household appliances based on web services, which allows the users to carry out efficient data analysis and assist products design by providing curve modeling, laboratory environment modeling and performance modeling services. The architecture and realization of test platform have been illustrated and the scalable grid service has been described in details. To demonstrate the test platform based on web services, a prototype application has been constructed. The results show that the test platform is feasible and effective.

Adaptively Construct Banking Process with Tags Upon Services-oriented Grid

Li Qi1, Song Wu2

1Operation Centre, China Development Bank, Beijing, China2Services Computing Technology and System Lab

Cluster and Grid Computing Lab School of Computer Science and Technology

Huazhong University of Science and Technology, Wuhan, ChinaAbstract

In today’s highly competitive financial markets, financial innovation and bank’s merge and acquisition activities are growing much. How to quickly construct business process upon IT infrastructure becomes the kernel of competition. As one of the classic application of service-oriented architecture, the grid technology can considerably help banks to improve the efficiency of constructing business process. In this paper, a tag-based processing model upon service-oriented grid is proposed. Through the practical scenario, the efficiency and security are demonstrated. Compared with traditional methods, the proposed solution has three advantages:(i) it is general enough to apply to banking applications without modifying upper business logic; (ii) with minor cost on processing normal request, it can provide high security and facility for reconstruction of banking business; (iii) it can help bank construct the new business process in the mean time.

21

Achievement for Complicated Electromagnetic Environment Simulation Application Based on ChinaGrid

Min Zhang1,Haitao Jia2,Shunsheng Zhang2

1Information Center, University of Electronic Science and Technology of China2Research Institute of Electronic Science and TechnologyUniversity of Electronic Science and Technology of China

Abstract

Complicated Electromagnetic Environment Simulation still puzzles the researchers who are engaged in System Design of large-scale electron system. Even electromagnetic simulation for a less dielectric has great iterative numeration, and complicated environment will refer to more and more calculation units belong to intricate electron equipment and national element. Mass compute, data alternation transfers and complex electromagnetic mechanism actualization will let researcher to go mad. This paper presents a novel system framework to build a complicated electromagnetic environment simulation based on ChinaGrid, which is agile and distributed federations. It can achieve the goal of the simulation system and administer to system design for large electron system.

Designing a Chemical Script Editor in Grid environment with DLTKYongying He, Ruisheng Zhang, Ruipeng Wei, Caihua Hu, Lian Li

Engineering Research Center of Open Source Software and Real–time Systems, Ministry of Education

School of Information Science and Engineering, Lanzhou University Lanzhou, Gansu, China

Abstract

All chemistry software has its own language and provides its corresponding script editor. A complex chemical job running in Grid environment often requires different chemistry software’s collaboration. The chemist edits computational chemistry program locally, then submits them to the Grid environment. Therefore, chemists need to use appropriate chemical editors for syntax checking when modifying chemical scripts. This process would require installing several kinds of chemistry software locally, which would be time-consuming and inconsequence. This paper proposes a solution, a Chemical Script Editor supporting several chemical script languages, to deal with this issue. The Chemical Script Editor is not only multifunctional and multi-lingual but also simple and easy to install. What’s more, the Chemical Script Editor caters to the dynamic characteristics of Grid environment perfectly. This paper would use a new framework -DLTK- to generate the Chemical Script Editor as a PLUG-INS to Eclipse platform.

22

Realcouse2.0: From Fully Connected Topology to Multiple Overlapping Stars

Jinyu Zhang, Xiaoming LiNet Lab, School of EECS, Peking University, Beijing, China

Abstract

We present the second version of Realcourse. Realcourse is a cooperative Internet video publishing and delivery system used by more than thirty Chinese universities to share course videos. The design of Realcourse bears the following assumptions 1) the system is supposed to consists of hundreds of unreliable but controllable servers where temporal failures are common, permanent failures are rare, 2) videos are concurrently uploaded to local servers but their metadata are broadcasted to all servers, 3) users are connected to and served by local servers but should be able to watch all the videos in the system in a seamless and efficient fashion. Realcourse has gone through two versions. In Realcourse1.0, all servers were treated equal and fully connected by messaging channels. In order to maintain a consistent global directory among servers, each server informs all others of local update operations through the message channels. The fully connected server topology poses a great challenge to system management with respect to failures. To solve this problem, Realcourse2.0 adopts a server topology of multiple overlapping stars by 1) replacing direct broadcast with relayed-broadcast, in which message is first sent to a relay server and then forwarded to all other servers, 2) partitioning the global directory into sub-trees, each sub-tree is then fully replicated to a group of selected servers while groups may overlap with each other. This paper explains why such a change is necessary and how it was done. Realcourse has been operational since December 2003. Today, more than 5000 videos are served. We constantly observe about 10000 unique user IP addresses each day.

23

Session 6: Semantic Grid and Semantic Web

KeyOnto: A Hybrid Knowledge Retrieval Model in Law Semantic WebBiao Fan1,2, Guangqiang Liu1,2, Tao Liu1,2, He Hu1,2 and Xiaoyong Du1,2

1Key Laboratory of Data Engineering and Knowledge EngineeringMOE, Beijing, China

2School of Information, Renmin University of China, Beijing, ChinaAbstract

This paper proposes a hybrid knowledge retrieval model KeyOnto, which combines ontology based knowledge retrieval model with traditional Vector Space Model (VSM). KeyOnto model makes use of domain ontology to organize and structure knowledge resources. Documents and queries are represented by concepts and term vectors respectively. Furthermore, ontology based query expansion called K2CM, is introduced to get expanded concepts of a query. Domain specific terms are used to form a term vector for queries and documents. Basing on these vectors, we can evaluate term similarity and concept similarity respectively, and integrate them together. Domain specific thesaurus is used to assist knowledge retrieval. Experiments show that compared with each single model, KeyOnto model improves precision of query result.

Ontology Based Semantic Relation Verification for TCM Semantic GridXiaogang Zhang,Huajun Chen, Jun Ma, Jinhuo Tao

Department of Computer Science, Zhejiang University, Hang Zhou, ChinaAbstract

Traditional Chinese Medicine (TCM) Semantic Grid is an application of Semantic Grid technique in TCM domain. It comprises TCM ontology and TCM Databases as data resources and Dart Search, Dart Query etc. as TCM application. This paper reports an ontology engineering component of the TCM Semantic Grid. Although ontology engineering has gained significant progress, ontology construction still mainly depends on manual work and tends to be mistaken prone. In this paper, we introduce a semantic relation verification method based on both domain ontology and domain publications. A modified vector space model is used to extract semantic relations from domain publications, which is particularly useful when the semantic relation cannot be extracted directly. Association rule learning method is used to distinguish significant relations from trivial ones. Further verification method is used to give user recommendations of relation types. We use Traditional Chinese Medicine Language System, domain ontology for Traditional Chinese Medicine, and relevant publications to validate our approach. But our method is not limited to this field. In fact, any data source that can be extracted into relevant instance pairs is applicable.

24

Research of Ontology Modeling in Structure Engineering GridLONG hao1,2, LU Hai1, Di Rui-Hua1

1School of Computer Science,Beijing University of Technology,Beijing, China 2School of Software,JiangXi Normal University,Nanchang, China

Abstract

Traditional Chinese Medicine (TCM) Semantic Grid is an application of Semantic Grid technique in TCM domain. It comprises TCM ontology and TCM Databases as data resources and Dart Search, Dart Query etc. as TCM application. This paper reports an ontology engineering component of the TCM Semantic Grid. Although ontology engineering has gained significant progress, ontology construction still mainly depends on manual work and tends to be mistaken prone. In this paper, we introduce a semantic relation verification method based on both domain ontology and domain publications. A modified vector space model is used to extract semantic relations from domain publications, which is particularly useful when the semantic relation cannot be extracted directly. Association rule learning method is used to distinguish significant relations from trivial ones. Further verification method is used to give user recommendations of relation types. We use Traditional Chinese Medicine Language System, domain ontology for Traditional Chinese Medicine, and relevant publications to validate our approach. But our method is not limited to this field. In fact, any data source that can be extracted into relevant instance pairs is applicable.

25

Session 7: Cluster Computing and Virtual Machine

Optimize the Performance of Virtual Machine Checkpointing via Memory Exclusion

Haikun Liu, Hai Jin, Xiaofei LiaoServices Computing Technology and System Lab

Cluster and Grid Computing Lab School of Computer Science and Technology

Huazhong University of Science and Technology, Wuhan, China

Abstract

Virtual Machine (VM) level checkpoints bring several advantages which process-level checkpoint implementation can hardly provide: compatibility, transparence, flexibility, and simplicity. However, the size of VM-level checkpoint may be very large and even in the order of gigabytes. This disadvantage causes the VM checkpointing and restart time become very long. To reduce the size of VM checkpoint, this paper proposes a memory exclusion scheme using ballooning mechanism, which omits saving unnecessary free pages in the VM. We implement our prototype in Xen environment. Experimental measurements show our approach can significantly reduce the size of VM checkpoint with minimal runtime overhead, thereby greatly improve the checkpoint performance.

A Reliable Parallel Interval Global Optimization Algorithm Based On Mind Evolutionary Computation

Yongmei Lei,Shaojun ChenSchool of Computer Engineering and Science

Shanghai University, Shanghai , China

Abstract

In this paper, we investigate the parallel reliable computational model and propose a parallel interval evolutionary algorithm that integrates interval arithmetic and Mind Evolutionary Computation method. The major aim is to explorer the new parallel interval decomposition scheme can solve computation intensive problem and can determine the all optimal solution reliably. The proposed algorithm is experimentally testified on the ZiQiang 3000 cluster of Shanghai High Education Grid-e-Grid Computational Application Platform with a test suit containing 6 complex multi-modal function optimization benchmarks.

26

Correctness Analysis based on Testing and Checking for OpenMP Programs

Jianjiang Li, Dan Hei, Lin YanDepartment of Computer Science and Technology

University of Science and Technology Beijing, Beijing, ChinaAbstract

Using OpenMP to develop multithreaded programs is very simple and easy. However, wrong execution results occur when directives are not properly used. In this paper, a correctness testing and checking technique based on online-offline mixed mode is proposed. Where, online correctness testing only needs to record the data relative to those parallel regions of which execution results are not correct, and at the same time, it is able to identify all wrong parallel regions through one -parse online correctness testing. Offline correctness testing needn’t record relative data of parallel regions and is able to flexibly test the correctness of parallel regions modified repeatedly. The technique overcomes the shortcomings of the high overheads of the existing correctness tools which can’t repeatedly test the correctness of paral lel regions.

Research on The Performance of xVM Virtual Machine Based on HPCCTiezhu Zhao1, Yilong Ding1, Verdi March2,3, Shoubin Dong1, Simon See2,4

1Guangdong Key Laboratory of Computer NetworkSouth China University of Technology, Guangzhou, China

2Asia-Pacific Science and Technology Center (APSTC), Sun Microsystems 3Department of Computer Science, National University of Singapore

4Department of Mechanical & Aerospace EngineeringNanyang Technological University

AbstractThe virtual machine (VM) technology has received an increasing interest a

spotlight both in the industry and the research communities. Although the potential advantages of virtualization in HPC workloads have been documented, the potential impact to application performance in HPC environments is not clearly understood. This paper presents a study on performance evaluation of virtual HPC systems using High Performance Computing Challenge (HPCC) benchmark suite and xVM as the workload representative and VM technology, respectively. Based on the extended AHP (Analytic Hierarchy Process) method, we propose an efficient performance evaluation model based on extended AHP and analyze the results and quantify the performance overhead of xVM in terms of compute, memory, and network overhead. Our analysis shows that the computational and network performance in HVM is slightly better and the memory performance is significantly better compared to paravirtualization.

27

The Application of Virtual Machine on System Security

Siqin Zhao,Kang Chen,Weimin ZhengDepartment of Computer Science and Technology

Tsinghua University, Beijing, China

Abstract

It is very important to protect critical resources such as private data and code in computer systems. It is promising to protect private data and to improve the system security by leveraging the isolation attribute of virtual machine (VM). The isolation attribute of VM is provided by Virtual Machine Monitor (VMM) that runs in higher priority than guest OSes. If the critical components are isolated in VMs, access control can be enforced or attestation can be made when the subject is accessing via VMs. In this way, VMs can improve the security level of critical components such as OS, kernel, data, and applications. For computer system security, VMs can be used to detect malware intrusions and to protect critical components, which can be implemented by integrating detection or protection mechanism in either VMM or VMs. Authentication is required to create trustful VMs. This paper surveys technologies related of using virtual machines to enhance system security.

28

Session 8: Peer-to-Peer Computing

Event-Driven Scheduling for P2P VoDBin Cheng1, Hai Jin1, Xiaofei Liao1, Bo Li2

1Services Computing Technology and System Lab, Cluster and Grid Computing Lab, School of Computer Science and Technology, Huazhong

University of Science and Technology, China2Department of Computer Science, Hong Kong University of Science and

TechnologyAbstract

Scheduling is essential in Peer-to-Peer (P2P) Video-on-Demand (VoD) applications to obtain desirable user experience and high sharing efficiency. Existing systems and proposals largely utilize scheduling mechanisms that are periodical in nature. We refer these as time-driven scheduling. In this paper we argue that the time-driven scheduling algorithm is not efficient for P2P VoD due to its inadequate resource utilization and inability in quickly adapting to the potential dynamics in such systems. We propose an event-driven scheduling algorithm for P2P VoD systems, in which the scheduler is triggered by events such as peer churn or/and random seeks. We evaluate our design through extensive simulations, and the experimental results show that event-driven scheduling outperforms time-driven scheduling in terms of both user experience and system scalability.

Towards Data Revelation and Scheduling Policy in Mesh-based Peer-to-Peer Streaming

Qingchao Cai, Xuejie ZhangSchool of Information Science and Engineering, Yunnan University, Kunming,

ChinaAbstract

The data revelation policy in traditional mesh-based peer-to-peer streaming protocols always make peers announce all available data to their neighbors. However, we find that, in a low bandwidth network environment, this policy suffers from serious data loss. In addition, only data deadline and bandwidth are considered in the data scheduling algorithms in the existed protocols, which leads to the slow dissemination of newly generated content data.

This paper first presents a strategic segment revelation policy that the source under-report the available data to its neighbors, while other peers announce all available data. According to simulation results, this policy can accelerate the dissemination of new segments, especially in low bandwidth environment. Then we propose a simple model of data scheduling which reveals the relation of download rate and other key factors in data scheduling. Based on this model, we further propose a greedy data scheduling algorithm, each data segment is given a priority according to its deadline and distribution among neighbors, and the segments with higher priority would be earlier requested such that these segments can be downloaded in the shortest

29

time. The simulation results also indicate that the greedy data scheduling algorithm performs near-optimally in terms of bandwidth utilization.

30

Session 9: Cloud Computing

MReC4.5: C4.5 ensemble classification with MapReduceGongqing Wu1, Haiguang Li1, Xuegang Hu1, Yuanjun Bi1, Jing Zhang1,

Xindong Wu1,2

1 School of Computer Science and Information Engineering Hefei University of Technology, Hefei, China

2 Department of Computer Science, University of Vermont Burlington, U.S.A.Abstract

Classification is a significant technique in data mining research and applications. C4.5 is a widely used classification method, and ensemble learning adopts a parallel and distributed computing model for classification. Based on analyses of the MapReduce computing paradigm and the process of ensemble learning, we find that the parallel and distributed computing model in MapReduce is appropriate for implementing ensemble learning. This paper takes the advantages of C4.5, ensemble learning and the MapReduce computing model, and proposes a new method MReC4.5 for parallel and distributed ensemble classification. Our experimental results show that increasing the number of nodes would benefit the effectiveness of classification modeling, and serialization operations at the model level make the MReC4.5 classifier “construct once, use anywhere”.

RAS-M:Resource Allocation Strategy based on Market Mechanism in Cloud Computing

Xindong YOU, Xianghua XU, Jian WAN, Dongjin YUSchool of Computer Science and Technology Hangzhou Dianzi University, Hangzhou, China

AbstractResource management is one of the main issues in Cloud Computing. In order to

improve resource utilization of large Data Centers while delivering services with higher QoS to Cloud Clients, a resource allocation strategy based on market (RAS-M) is proposed. Firstly, the architecture and the market model of RAS-M are constructed, in which a QoS-refection utility function is designed according to different resource requirements of the Cloud Client, the equilibrium state of RAS-M is defined and the proof of its optimality is given. Secondly, GA-based price adjusted algorithm is introduced to deal with the problem of achieving the equilibrium state of RAS-M. Finally, RAS-M is implemented upon Xen to reallocate the VM’s weight. Experiments results obtained by setting different parameters show that RAS-M can achieve the equilibrium state approximately, that is, demand and supply is balanced

31

nearly, which validates RAS-M is effective and practicable, and is capable of achieving its goal.

32