Intelligent Cache Farming Architecture for network management

Abstract

Web caching is widely recognized as an effective technique that improves the quality of service (QoS) over the Internet, such as reducing user latency and network bandwidth usage. However, it still meets problems when the number of users increases due to limitations of hardware and management policy of caches. Therefore, this research proposes the Intelligent Cache Farming Architecture (ICFA) to be an alternative caching architecture model integrating with the recommending mechanism. The proposed architecture is the cache grouping mechanism where browsing characteristics are applied to improve the performance of the Internet services. In order to prove the proposed architecture, the trace-driven simulation and the traditional caching model, which is the original model without grouping criteria for the web cache, are implemented in a virtual machine environment. The results indicate that the delay measurement of the ICFA is dropped more than 44%, while the hit rate and the byte hit rate of the ICFA increase more than 20% and 43%, respectively.