Design Issues and Performance Analysis of Layered Web Cache Replacement Policy on the Internet

Abstract

The network traffic flow has more increasing due to more HTTP requests on the Internet. This growth may result in traffic congestion and heavy server load on the Internet. Hence, it becomes crucial to provide solutions for reducing the Web accessing latency. In addition to increasing the network bandwidth, Web caching is one of the most essential solutions that can fix these problems. There are many benefits of Web caching on proxy, including decreasing network traffic, reducing latency in fetching Web pages, and balancing I/O loads on heavy Web servers.
When the requested objects cannot be found in the Web cache or Web cache is full, the removed objects are determined by the efficient Cache Replacement policy to achieve fewer removals and higher cache utilization. In this paper, we employ layered concept into filtering, maintenance, and replacement mechanisms to design a new and efficient Cache Replacement policy called LWCRP. Then, we propose a strategy based upon trace-driven simulation to evaluate and compare performance between our policy and other conventional policies. Finally, we claim that our new policy can improve the processing performance and enhance the Quality of Service (QoS) on the Web servers.