Using Layered Bottlenecks for Virtual Machine Provisioning in the Clouds
UCC '12 Proceedings of the 2012 IEEE/ACM Fifth International Conference on Utility and Cloud Computing
| Hi-index | 0.00 |
High availability, reliability and scalability are basic prerequisites for cloud applications. Due to dynamically varying workloads, it's necessary to provide resource guarantees to cloud applications for meeting QoS requirements. However, it's not trivial to generate a precise scalability policy for multi-tiers cloud applications to adapt to dynamically varying workload and satisfy QoS requirements simultaneously with minimal resource consumption. In this paper, we present an on-line scalability controller of multi-tiers applications to handle with unpredictable changing workloads. The solution is based on the use of layered queuing network model combined with our dynamic resource allocation techniques. Our performance model relates application performance to their resource requirements and dynamically varying workloads. Then we evaluate our approach via experiments with real-world workloads in different scenarios. Our results indicate that our performance model faithfully captures the behaviors of multi-tiers applications over a various range of workloads and configuration schemes. Moreover, it also shows that our techniques can judiciously obtain the optimized configuration scheme effectively with modest computation.