원문정보
초록
영어
As more and more Cloud systems are designed and deployed in business and science engineering areas, how to evaluate and monitor the system’s runtime performance is becoming an important issue. However, few efforts are taken to address such an issue because the resources in Cloud systems often owned by different institutes. In this paper, we present an integrated performance evaluation framework, which is aiming to provide cloud users an easy-to-use toolkit to evaluate their system’s runtime performance, or compare the performance under different resource management policies. The design and implementation of our framework is highly extensible and re-useable in most existing cloud systems. Currently, the prototype of our implementation is examined by a series of experiments, and the results indicate that its configurable feature is very useful when users are conducting performance comparing under different contexts.
목차
1. Introduction
2. Related Work
3. System Framework
3.1. Architecture
3.2. Evaluation Procedure
3.3. Performance Profiling Model
4. Experiments and Performance Evaluation
4.1. Effects of VM Management Policy on Performance
4.2. Effects of Scheduling Algorithm on Performance
5. Conclusion
Acknowledgements
References