Maximizing Information Freshness in Caching Systems with Limited Cache Storage Capacity

TL;DR

This paper analyzes how to optimally manage limited cache storage to maximize information freshness at a user, balancing between caching files and fetching directly from the source considering update rates and transmission times.

Contribution

It provides analytical expressions for file freshness and determines optimal caching strategies and update rates to maximize overall information freshness.

Findings

Caching improves freshness when cache update rates are high.

Direct source access is preferable for frequently changing files with low transmission times.

Optimal policies depend on update rates and transmission time trade-offs.

Abstract

We consider a cache updating system with a source, a cache with limited storage capacity and a user. There are $n$ files. The source keeps the freshest versions of the files which are updated with known rates. The cache gets fresh files from the source, but it can only store the latest downloaded versions of $K$ files where $K\leq n$. The user gets the files either from the cache or from the source. If the user gets the files from the cache, the received files might be outdated depending on the file status at the source. If the user gets the files directly from the source, then the received files are always fresh, but the extra transmission times between the source and the user decreases the freshness at the user. Thus, we study the trade-off between storing the files at the cache and directly obtaining the files from the source at the expense of additional transmission times. We find analytical expressions for the average freshness of the files at the user for both of these scenarios. Then, we find the optimal caching status for each file (i.e., whether to store the file at the cache or not) and the corresponding file update rates at the cache to maximize the overall freshness at the user. We observe that when the total update rate of the cache is high, caching files improves the freshness at the user. However, when the total update rate of the cache is low, the optimal policy for the user is to obtain the frequently changing files and the files that have relatively small transmission times directly from the source.