2018/06/28 by Eric M. Friedlander, Friedlander, Eric, Vaneet Aggarwal +1
Computer Science · Engineering · #Advanced Wireless Network Optimization #Caching and Content Delivery #Distributed #FOS: Computer and information sciences #Multimedia (cs.MM) #Networking and Internet Architecture (cs.NI) #Parallel #Peer-to-Peer Network Technologies #and Cluster Computing (cs.DC)
paper · pdf · doi:10.48550/arxiv.1806.10853
openalex publication_date 2018/06/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Caching plays a crucial role in networking systems to reduce the load on the\nnetwork and is commonly employed by content delivery networks (CDNs) in order\nto improve performance. One of the commonly used mechanisms, Least Recently\nUsed (LRU), works well for identical file sizes. However, for asymmetric file\nsizes, the performance deteriorates. This paper proposes an adaptation to the\nLRU strategy, called gLRU, where the file is sub-divided into equal-sized\nchunks. In this strategy, a chunk of the newly requested file is added in the\ncache, and a chunk of the least-recently-used file is removed from the cache.\nEven though approximate analysis for the hit rate has been studied for LRU, the\nanalysis does not extend to gLRU since the metric of interest is no longer the\nhit rate as the cache has partial files. This paper provides a novel\napproximation analysis for this policy where the cache may have partial file\ncontents. The approximation approach is validated by simulations. Further, gLRU\noutperforms the LRU strategy for a Zipf file popularity distribution and\ncensored Pareto file size distribution for the file download times. Video\nstreaming applications can further use the partial cache contents to help the\nstall duration significantly, and the numerical results indicate significant\nimprovements (32 %) in stall duration using the gLRU strategy as compared to\nthe LRU strategy. Furthermore, the gLRU replacement policy compares favorably\nto two other cache replacement policies when simulated on MSR Cambridge Traces\nobtained from the SNIA IOTTA repository.\n