A Server Cache Size Aware Cache Replacement Algorithm for Block Level Network Storage

Abstract

Storage management cost is one of the most important issues in the current computer system. Network storage is deployed in many companies to decrease this cost. In the network storage system, both the server computer and network storage have cache memory for the same storage. In most cache implementation, LRU replacement algorithm, which is suitable for accesses with temporal locality of reference, is adopted. However, accesses to network storage do not have usual temporal locality but have negative temporal locality. Consequently, caches in network storages do not work effectively. In this paper, we propose a new cache replacement algorithm, which is suitable for negative temporal locality of reference. Our simulation evaluation demonstrated that the proposed algorithm provides better cache hit ratio than other existing algorithms.