Weighted Greedy Dual Size Frequency Based Caching Replacement Algorithm

Tinghuai Ma,Jingjing Qu,Wenhai Shen,Yuan Tian,Abdullah Al-Dhelaan,Mznah Al-Rodhaan
DOI: https://doi.org/10.1109/access.2018.2790381
IF: 3.9
2018-01-01
IEEE Access
Abstract:Caches are used to improve the performance of the internet, and to reduce the latency of data access time and the low speed of repeated computing processes. Cache replacement is one of the most important issues in a caching system; therefore, it must be coordinated with the caching system to minimize the access latency and maximize the hit rate or byte hit rate. In this paper, we presented a novel caching replacement algorithm named Weighted Greedy Dual Size Frequency (WGDSF) algorithm, which is an improvement on the Greedy Dual Size Frequency (GDSF) algorithm. The WGDSF algorithm mainly adds weighted frequency-based time and weighted document type to GDSF. By increasing the above two weighted parameters, WGDSF performs fairly well at keeping popular objects in the cache and replacing rarely used ones. Our experiment shows that this algorithm has a better hit rate, byte hit rate and access latency than state-of-the-art algorithms, such least Recently Used, least Frequently Used, and GDSF.
What problem does this paper attempt to address?