Modeling and performance analysis of large scale IaaS Clouds
Future Generation Computer Systems
Cloud Computing Operations Research
Service Science
| Hi-index | 0.00 |
Optimizing for performance is often associated with higher costs in terms of capacity, faster infrastructure, and power costs. In this paper, we quantify the power-performance trade-offs by developing a scalable analytic model for joint analysis of performance and power consumption for a class of Infrastructure-as-a-Service (IaaS) clouds with tiered service offerings. The tiered service offerings are provided by configuring physical machines into three pools with different response time and power consumption characteristics. Using interacting stochastic sub-models approach, we quantify power-performance trade-offs. We summarize our modeling approach and highlight key results on the effects of physical machine pool configurations on consumed power and achievable performance in terms of response time and ability to service requests. The approach developed here can be used to manage power consumption and performance by judiciously configuring physical machine pools.