Critical Sizing of LRU Caches with Dependent Requests

It was recently proved in [12] that the Least-Recently-Used (LRU) caching policy, in the presence of semi-Markov modulated requests that have a generalized Zipf’s law popularity distribution, is asymptotically insensitive to the correlation in the request process. However, since the preceding result is asymptotic, it remains unclear how small the cache size can become while still retaining this insensitivity property. In this paper, assuming requests come from a nearly completely decomposable Markov-modulated process, we characterize the critical cache size below which the dependency of requests dominates the cache performance. It appears surprising that the critical cache size is very small, and in fact it is sub-linear to the sojourn times of the modulated chain that determines the dependency structure.

By: Predrag Jelenkovic; Ana Radovanovic; Mark S. Squillante

Published in: RC23779 in 2005


This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.


Questions about this service can be mailed to .