Optimal Load Balancing in Distributed Computer Systems

ISBN-13: 9781447112464 / Angielski / Miękka / 2011 / 251 str.

Hisao Kameda; Jie Li; Chonggun Kim

Opis

An important consideration in improving the performance of a distributed computer system is the balancing of the load between the host computers. Load balancing may be either static or dynamic; static balancing strategies are generally based on information about the system's average behavior rather than its actual current state, while dynamic strategies react to the current state when making transfer decisions. Although it is often conjectured that dynamic load balancing outperforms static, careful investigation shows that this view is not always valid. Recent research on the problem of optimal static load balancing is clearly and intuitively presented, with coverage of distributed computer system models, problem formulation in load balancing, and effective algorithms for implementing optimization. Providing a thorough understanding of both static and dynamic strategies, this book will be of interest to all researchers and practitioners working to optimize performance in distributed computer systems.

Spis treści

1 Single Channel and Star Network Configurations.- 1.1 Introduction.- 1.2 Load Balancing in the Single Job Class Environment.- 1.2.1 Introduction.- 1.2.2 Model description.- 1.2.3 Properties of the optimal solution and an optimal load balancing algorithm.- 1.2.4 Comparison of load balancing algorithm performance.- 1.2.5 An algorithm for star network configurations.- 1.2.6 Conclusion.- 1.3 Load Balancing in the Multi-class Job Environment.- 1.3.1 Introduction.- 1.3.2 Model description.- 1.3.3 Optimal solution.- 1.3.4 Optimal load balancing algorithm.- 1.3.5 Comparison of algorithm performance.- 1.3.6 Conclusion.- 2 Overall vs. Individual Optimum.- 2.1 Introduction.- 2.2 Single Channel Communications Networks.- 2.2.1 Model Description.- 2.2.2 Optimal Solutions.- 2.2.3 Parametric Analysis.- 2.2.4 Anomalous Behaviors of the Optimum and the Equilibrium.- 2.2.5 Numerical Examination.- 2.2.6 Conclusion.- 2.3 Star Network Configurations.- 2.3.1 Model Description.- 2.3.2 Optimal Solutions.- 2.3.3 Parametric Analysis.- 2.3.4 Anomalous Behaviors of the Performance Variables.- 2.3.5 Numerical Examination.- 2.3.6 Discussion.- 2.3.7 Conclusion.- 2.4 Multiclass Single Channel Networks.- 2.4.1 Policies and model.- 2.4.2 Numerical experiment.- 2.4.3 Discussion.- 2.4.4 Conclusion.- 3 Tree Hierarchy Network Configurations.- 3.1 Introduction.- 3.2 Model Description and Problem Formulation.- 3.3 Optimal Load Balancing.- 3.4 Decomposability.- 3.5 Proposed Algorithm.- 3.6 Comparison of Algorithm Performance.- 3.6.1 Comparison of Storage Requirements.- 3.6.2 Comparison of Computation Time Requirements.- 3.7 Link Communication Time and Node Processing Time.- 3.7.1 Concept of Sub-Tree Networks.- 3.7.2 Parametric Analysis.- 3.7.3 Effects of Link Communication Time.- 3.7.4 Effects of Node Processing Time.- 3.8 Conclusion.- 4 Star Network with Two-way Traffic.- 4.1 Introduction.- 4.2 Model Description and Problem Formulation.- 4.3 Necessary and Sufficient Conditions.- 4.4 Proposed Algorithm.- 4.5 Parametric Analysis.- 4.6 A Numerical Example.- 4.7 Discussion.- 4.8 Conclusion.- 5 Tree Network with Two-way Traffic.- 5.1 Introduction.- 5.2 Model Description and Problem Formulation.- 5.3 Necessary and Sufficient Conditions.- 5.4 Decomposition.- 5.5 Proposed Algorithm.- 5.6 Comparison of Algorithm Performance.- 5.6.1 Comparison of Storage Requirements.- 5.6.2 Comparison of Computation Time Requirements.- 5.7 Conclusion.- 6 Uniqueness.- 6.1 Introduction.- 6.2 Description of the Model.- 6.3 The Overall Optimal Solution.- 6.4 The Individually Optimal Solution.- 6.5 Numerical Examples.- 6.6 Concluding Remarks.- 7 A Survey of Dynamic Load Balancing.- 8 Static vs. Dynamic.- 8.1 System Model.- 8.2 Static and Dynamic.- 8.2.1 Static Load Balancing Policies.- 8.2.2 Dynamic Load Balancing Policies.- 8.3 Simulation Results.- 8.4 Discussion.