Power Aware Hybrid Proxy Cache-Prefetch Model using Combined Energy Conservation Policies

The World Wide Web (WWW) is growing exponentially in terms of number of users and number of Web applications. Due to enormous traffic in the network and several factors like bandwidth availability, request processing time at server, round trip time and object size, the Web latency is increasing. The sophisticated integration of Web prefetching and caching deployed at proxy server with Web log mining technique is the most attractive and successful solution to reduce this latency and improve Web Quality of Service. To provide uninterrupted services to the Web users’, Web servers at the data centres are seamlessly operating always in 24X7 mode. Low power consumption and efficient energy management of individual hardware components, system software, to network applications is a critical and urgent issue in today’s highly demanding eco-friendly computing world. In this paper, we incorporate an energy efficient feedback driven control framework along with sleep proxy mechanism in our Hybrid Cache-Prefetch System Model. The proposed model periodically monitors proxy server performance parameters dynamically. We have also developed an optimized load allocation algorithm with the proposed combined strategy for proxy servers’ cluster along with some guidelines for energy efficient design solutions towards green IT. Even though we emphasize power saving as much as possible the performance of proxy servers is ensured without sacrificing on client experiences. Experimental results show that using our proposed scheme 28% less power consumption and 18% power efficiency improvement can be achieved. Keywords— Proxy Cache-Prefetch model, Power and Energy Estimation, Feedback Driven Control Framework, Sleep Proxy, Optimal Load Allocation.