期刊名称:International Journal of Computer Science and Network Security
印刷版ISSN:1738-7906
出版年度:2011
卷号:11
期号:10
页码:64-69
出版社:International Journal of Computer Science and Network Security
摘要:In this article, we present a novel media caching algorithm adequate for real-world streaming workload. Streaming workload is believed to exhibit very large and sequential access characteristics, which is the main concern of a legacy media caching policy. However, a legacy caching algorithm did not fully utilize access pattern and also tends to overlook streaming interactivity which can decrease the advantage of a caching technique utilizing sequential reference. As a replaced cache is determined based on the expected caching gain of the candidates, cache hit ratio can be significantly improved. We develop Least Expectation First technique which can manage very large number of block caches as two-level grouping. Experimental results show that the proposed media caching scheme yields better hit ratio and I/O smoothing than legacy buffer cache replacement schemes.