Probability Models for Multiprogramming Computer Systems
Journal of the ACM (JACM)
Performance Evaluation and Monitoring
ACM Computing Surveys (CSUR)
On the construction of a representative synthetic workload
Communications of the ACM
The role of computer system models in performance evaluation
Communications of the ACM
System performance evaluation: survey and appraisal
Communications of the ACM
Further analysis of a computing center environment
Communications of the ACM
Determining a computing center environment
Communications of the ACM
| Hi-index | 0.00 |
A macro-model for evaluation of a large computer system is described. Estimates of system capacity and a simulated job load are used to compute a metric for the fraction of capacity which is utilized. This, coupled with estimates of average turnaround time, permits analysis of the “throughput/service” trade-off. The interrelationships of system measurement parameters are emphasized, and the large computing facility is characterized as a multiproduct firm which makes joint products available to its patrons. A simulation model which is under development is outlined, including examples. The simulations described herein consider two joint variables—central processor time and input/output activity (bytes transferred). Although the model can accommodate additional variables, the two variable case is sufficient for demonstrating the folly of failing to recognize the interaction of system parameters in determining the capacity of the system. Proper treatment of interdependent system parameters (joint products) in performance evaluation is the main thrust of the paper.