Evaluation of load sharing in locally distributed systems
Load sharing has been the focus of a great deal of research as a means of enhancing the performance of distributed systems. This dissertation evaluates the potential benefits of load sharing when important system characteristics are varied. The author starts with a baseline model that consists of homogeneous, monolithic nodes executing independent jobs. The first variation to the baseline model allows heterogeneous nodes, i.e., nodes with different speeds. Using Markov decision theory, it is shown that, except in highly heterogeneous systems, the queue length scaled by node speed yields an almost optimal policy. The second variation deals with cluster-based distributed systems. It is shown that most of the benefits can be realized by employing load sharing within each cluster alone. The third variation considers multiple-resource nodes. The model captures systems with local I/O devices. It was found that the performance gain from load sharing diminishes with node complexity regardless of routing, utilization, system size, or the load measure used. Finally, the nature of workload is changed to combine the two views of task allocation and job scheduling traditionally treated separately. Each job may have multiple concurrent modules.
- Research Organization:
- Washington Univ., Seattle (USA)
- OSTI ID:
- 6565957
- Resource Relation:
- Other Information: Thesis (Ph. D.)
- Country of Publication:
- United States
- Language:
- English
Similar Records
Multi-core processing and scheduling performance in CMS
D-Factor: A Quantitative Model of Application Slow-Down in Multi-Resource Shared Systems
Related Subjects
DISTRIBUTED DATA PROCESSING
TASK SCHEDULING
ARRAY PROCESSORS
FUNCTIONAL MODELS
MARKOV PROCESS
PERFORMANCE
QUEUES
DATA PROCESSING
PROCESSING
STOCHASTIC PROCESSES
990300* - Information Handling
990210 - Supercomputers- (1987-1989)