Distributed and Parallel Systems (eBook)

Cluster and Grid Computing

Zoltan Juhasz, Peter Kacsuk, Dieter Kranzlmuller (Herausgeber)

eBook Download: PDF

2005 | 2005
XI, 211 Seiten
Springer US (Verlag)
978-0-387-23096-2 (ISBN)

Infos

DAPSY (Austrian-Hungarian Workshop on Distributed and Parallel Systems) is an international conference series with biannual events dedicated to all aspects of distributed and parallel computing. DAPSY started under a different name in 1992 (Sopron, Hungary) as regional meeting of Austrian and Hungarian researchers focusing on transputer-related parallel computing; a hot research topic of that time. A second workshop followed in 1994 (Budapest, Hungary). As transputers became history, the scope of the workshop widened to include parallel and distributed systems in general and the 1st DAPSYS in 1996 (Miskolc, Hungary) reflected the results of these changes.

Distributed and Parallel Systems: Cluster and Grid Computing is an edited volume based on DAPSYS, 2004, the 5th Austrian-Hungarian Workshop on Distributed and Parallel Systems. The workshop was held in conjunction with EuroPVM/MPI-2004, Budapest, Hungary September 19-22, 2004.

Inhalt

Contents 6
Preface 10
I GRID SYSTEMS 13
GLOGIN - INTERACTIVE CONNECTIVITY FOR THE GRID* 15
1. Introduction 15
2. Overview of Approach 16
Limitations of Globus-Gatekeeper 16
Getting Interactive Connections 17
3. Details of the Implementation 18
Connection Establishment 18
Secure Connection Establishment 19
Getting shells and other commands 20
Traffic forwarding 20
4. Related Work 21
5. Conclusions and Future Work 22
Acknowledgments 22
Notes 22
References 23
PARALLEL PROGRAM EXECUTION SUPPORT IN THE JGRID SYSTEM* 25
1. Introduction 25
2. Execution Support for the Grid 26
3. Parallel execution support in JGrid 26
The Batch Execution Service 27
The Compute Service 30
4. Results 32
5. Conclusions and Future Work 32
References 32
VL-E: APPROACHES TO DESIGN A GRID-BASED VIRTUAL LABORATORY 33
Introduction 33
1. The Virtual Laboratory Architecture 34
2. The concept of study in VL-E 36
3. Resource management in VL-E 36
4. Related Work 38
5. Conclusions 39
References 40
SCHEDULING AND RESOURCE BROKERING WITHIN THE GRID VISUALIZATION KERNEL* 41
1. Introduction 41
2. Related Work 42
3. The GVK Visualization Planner 43
4. Visualization Task Decomposition 44
5. Resource Information Gathering 44
6. Algorithm Selection and Resource Mapping 45
7. Visualization Pipeline Construction 46
8. Conclusions and Future Work 47
References 47
II CLUSTER TECHNOLOGY 49
MESSAGE PASSING VS. VIRTUAL SHARED MEMORY A PERFORMANCE COMPARISON 51
1. Introduction 51
2. The Virtual Shared Memory Paradigm 53
3. Benchmarks 54
Benchmark 1: Approximation 54
Benchmark 2: Tree Structured Matrix Multiplications 55
Benchmark 3: Eigenvector Accumulation 55
4. Experimental Results 55
5. Conclusion and Future Work 57
References 58
MPI-I/O WITH A SHARED FILE POINTER USING A PARALLEL VIRTUAL FILE SYSTEM IN REMOTE I/ O OPERATIONS 59
1. Introduction 59
2. Implementation of PVFS in Stampi 60
3. Performance measurement 62
Inter-machine data transfer 63
Local I/O operations 63
Remote MPI-I/O operations 64
4. Summary 65
Acknowledgments 66
References 66
AN APPROACH TOWARD MPI APPLICATIONS IN 67
WIRELESS NETWORKS 67
1. Introduction 67
2. Reviewing the Fault Detection Mechanism 68
3. Unconstrained Global Optimization for n-Dimensional Functions 69
Parallel Program Without Wireless Channel State Detection 69
Parallel Program With Wireless Channel State Detection 71
Experimental Results 72
4. Conclusions and Future Work 73
References 74
DEPLOYING APPLICATIONS IN MULTI- SAN SMP CLUSTERS 75
1. Introduction 75
2. Our Approach 76
3. Representation of Resources 76
Basic Organization 77
Virtual Views 78
4. Application Modelling 78
Entities for Application Design 78
A Modelling Example 79
5. Mapping Logical into Physical Resources 80
Laying Out Logical Resources 80
Dynamic Creation of Resources 81
6. Discussion 82
Notes 82
References 82
III PROGRAMMING TOOLS 83
EXAMPLES OF MONITORING AND PROGRAM ANALYSIS ACTIVITIES WITH DEWIZ 85
1. Introduction 85
2. Overview of 86
3. Analysis of OpenMP and PVM Pograms with DEWIZ 87
4. User-defined Visualization of Event-Graphs using the Client 89
5. Conclusions and Future Work 92
References 93
INTEGRATION OF FORMAL VERIFICATION AND DEBUGGING METHODS IN P- GRADE ENVIRONMENT* 95
1. Introduction to P- GRADE and DIWIDE 95
2. Coloured Petri- net and Occurrence Graph 98
3. Transformation steps from GRAPNEL to CPN 98
4. Steering the macrostep debugger based on simulation 100
5. Related works 103
6. Summary, future goals 103
References 104
TOOLS FOR SCALABLE PARALLEL PROGRAM ANALYSIS - VAMPIR NG AND DEWIZ 105
1. Introduction 105
2. Tools for Parallel Program Analysis 106
3. High- Performance Program Analysis with VNG 108
4. Distributed Program Analysis with DeWiz 109
5. Comparison and Insights 110
6. Summary and Future Work 112
Acknowledgments 113
Notes 113
References 113
PROCESS MIGRATION IN CLUSTERS AND CLUSTER GRIDS * 115
1. Introduction 115
2. The Hungarian ClusterGrid Project 116
3. The P-GRADE software development tool 117
4. Migration in the 1st generation ClusterGrid 118
5. Migration in the 2nd generation ClusterGrid 120
6. Performance and Related Work 121
7. Conclusion 121
References 122
IV P-GRADE 123
GRAPHICAL DESIGN OF PARALLEL PROGRAMS WITH CONTROL BASED ON GLOBAL APPLICATION STATES USING AN EXTENDED P- GRADE SYSTEM 125
1. INTRODUCTION 125
2. PS-GRADE - SYNCHRONIZATION -ORIENTED P-GRADE SYSTEM 126
3. EXAMPLE: A TSP SOLVED BY B& B METHOD
4. CONCLUSIONS 131
5. BIBLIOGRAPHY 132
PARALLELIZATION OF A QUANTUM SCATTERING CODE USING P- GRADE: A CASE STUDY 133
1. Introduction 133
2. Re-structuring of the FORTRAN code 134
3. Setting up the parallel code using P-GRADE 136
4. Program performance analysis 137
5. Summary 139
Acknowledgments 139
References 140
TRAFFIC SIMULATION IN P-GRADE AS A GRID SERVICE 141
1. Introduction 141
2. Traffic simulation using P-Grade 142
3. Grid Execution Management for Legacy Code Architecture 143
4. Integrating GEMLCA with the P-Grade portal 145
5. Conclusion 146
Acknowledgments 147
References 148
DEVELOPMENT OF A GRID ENABLED CHEMISTRY APPLICATION 149
1. Introduction 149
2. Reaction-diffusion equations 150
3. Parallel implementation in P-GRADE 151
4. Performance results on non-dedicated cluster 152
5. Performance results in the Grid 154
6. Related works 155
7. Summary 155
References 156
V APPLICATIONS 157
SUPPORTING NATIVE APPLICATIONS IN WEBCOM- G 159
1. Introduction 159
Program Execution in WebCom-G 2. 161
Extraction 161
Annotation 164
3. Automatic Parallelization 164
4. Conclusions and Future Work 165
Acknowledgments 165
References 165
GRID SOLUTION FOR E-MARKETPLACES INTEGRATED WITH LOGISTICS 167
1. INTRODUCTION 167
2. INTEGRATING LOGISTICS INTO EMARKETPLACES 168
3. ROLE OF GRID COMPUTING IN EMARKETPLACES 171
4. GRID EXECUTION MANAGEMENT FOR LEGACY CODE ARCHITECTURE 172
5. GRID SERVICES BASED E-MARKETPLACE MODEL WITH GEMLCA 173
6. SUMMARY 174
7. ACKNOWLEDGEMENTS 175
8. REFERENCES 175
INCREMENTAL PLACEMENT OF NODES IN A LARGE- SCALE ADAPTIVE DISTRIBUTED MULTIMEDIA SERVER 177
1. Introduction 177
2. Related Work 178
3. The problem model 178
4. Incremental algorithm 180
5. Results 181
6. Conclusions and Further Work 183
References 183
COMPONENT BASED FLIGHT SIMULATION IN DIS SYSTEMS 185
Introduction 185
1. DIS system architecture 186
2. Component interaction model 187
Simulation object 188
Remote object interaction 191
Human operator 192
3. Summary 193
Notes 193
References 193
VI ALGORITHMS 195
MANAGEMENT OF COMMUNICATION ENVIRONMENTS FOR MINIMALLY SYNCHRONOUS PARALLEL ML 197
1. Introduction 197
2. Minimally Synchronous Parallel ML 198
3. Management of Communication Environments 200
4. Comparison to Related Work 202
5. Conclusions and Future Work 203
References 204
ANALYSIS OF THE MULTI-PHASE COPYING GARBAGE COLLECTION ALGORITHM 205
Introduction 205
1. Multi-Phase Copying Garbage Collection Algorithm 206
2. Analysis of the algorithm 207
Number of phases in the MC-GC algorithm 209
3. Conclusion 212
References 212
A CONCURRENT IMPLEMENTATION OF SIMULATED ANNEALING AND ITS APPLICATION TO THE VRPTW OPTIMIZATION PROBLEM 213
1. INTRODUCTION 213
2. SIMULATED ANNEALING 214
3. COMMUNICATION SCHEME OF CONCURRENT SIMULATED ANNEALING 215
3.1 Implementation of communication with synchronization at solution acceptance events 216
3.2 Tuning of the algorithm 217
4. EXPERIMENTAL RESULTS 219
4.1 VRPTW 219
5. IMPLEMENTATION 219
6. CONCLUSIONS 220
REFERENCES 221
Author Index 222

Textauszug

GRID SOLUTION FOR E-MARKETPLACES INTEGRATED WITH LOGISTICS (p. 155-156)

Abstract:  Electronic marketplaces are important facilitators of today’s e-business activities. Besides substantial advantages offered by these exchange sites, e-marketplaces do not work up to their full potential at the moment. This paper describes both a new business model and its technical implementation using Grid technology. A new, three-sided e-commerce model is suggested that integrates buyers, sellers and logistics service providers who all participate in the same negotiation process. To solve computation intensive optimisation tasks and to integrate back-office and marketplace applications, a Grid services based marketplace implementation model is outlined.

Key words: e-business, logistics, e-marketplace, Grid service, legacy code


1. INTRODUCTION

The evolution of business models and technological solutions advance together like intertwining spirals motivating and supporting each other. Business requirements drive information technology (IT) to find new tools and techniques that make businesses develop new needs again. Early electronic commerce – Electronic Data Interchange (EDI) - was started because telecommunication between computers facilitated a new relationship between businesses.

The rise of new business needs resulted in new communication standards and protocols. Real ecommerce was born out of the opportunity offered by the World Wide Web and triggered new IT researches again. After the unrealistic hype of the 90’s and the crises around 2000 e-commerce by now has entered the reality phase where efficiency drives the businesses, Internet usage adds value and increases the profitability of the companies. (Plankett Research 2004) A main target area of seeking business efficiency is supply chain management (SCM). Today a substantial part of supply chains are managed across the Internet still they contain a surprisingly high amount of inefficiencies. (Oliver at al. 2002) Both business and technology sides should be revised to find ways of improvement.

New e-commerce models might be considered and the latest information technology tools searched for to support them. The Grid concept has been created for solving computation intensive scientific problems, but the possibility of business applications was soon discovered. The convergence between Web services and Grid computing, that was triggered by the specification of OGSA (Open Grid Services Architecture) (Foster et al., 2002), resulted in even more intensive interest from large industry players towards Grid-based solutions. OGSA defines a Grid architecture that is based on Web service standards and protocols. As Web services are becoming more and more common in business applications, a Grid architecture based on SOAP (Simple Object Access Protocol) communication and WSDL (Web Services Description Language) service descriptions is the natural model to adopt in a business environment.

This paper would like to provide a step forward both in the fields of business and technology. As an answer to SCM problems a new, threesided e-commerce model is suggested that integrates buyers, sellers and logistics service providers in the same negotiation process. This marketplace helps trading partners to minimise their costs and increase their profit. The online optimisation requires large amounts of computation without any delay, which has focused attention on Grid technology. Following the direction set by Kacsukné (2004) and using the Grid-based e-marketplace architecture introduced by Kiss at al. (2004) this article gives a complex picture of the new e-marketplace model and its planned implementation.

Produkt-Details

• Erscheint lt. Verlag: 31.12.2005
• Reihe/Serie: The Springer International Series in Engineering and Computer Science
• Reihe/Serie: The Springer International Series in Engineering and Computer Science
• Zusatzinfo: XI, 211 p. 48 illus.
• Verlagsort: New York
• Sprache: englisch
• Themenwelt: Mathematik / Informatik ► Informatik ► Betriebssysteme / Server
• Themenwelt: Mathematik / Informatik ► Informatik ► Netzwerke
• Themenwelt: Informatik ► Weitere Themen ► Hardware
• Schlagworte: Debugging • Distributed Systems • grid computing • Kernel • Media Server • Monitor • Optimization • Scheduling • Tools
• ISBN-10: 0-387-23096-3 / 0387230963
• ISBN-13: 978-0-387-23096-2 / 9780387230962