Knapsack-model-based Schemes and WLB Algorithm for the Capacity and Efficiency Management of Web Cache
Anbo Xiang, Liang Xu, Baozhuang Niu · International Journal of Computers Communications & Control · 2013
Web cache refers to the temporary storage of web files/documents. In reality, a set of caches can be grouped into a cluster to improve the server system's performance. In this paper, to achieve the overall cluster efficiency, we propose a weighted load balancing (WLB) routing algorithm by considering both the cache capability and the content property to determine how to direct an arrival request to the right node. Based on Knapsack models, we characterize three new placement/replacement schemes for Web contents caching and then conduct the comparison based on WLB algorithm. We also compare WLB algorithm with two other widely used algorithms: Pure load balancing (PLB) algorithm and Round-Robin (RR) algorithm. Extensive simulation results show that the WLB algorithm works well under the examined cluster content placement/replacement schemes. It generally results in shorter response time and higher cache hit ratio, especially when the cache cluster capacity is scarce.