Cache-aided private information retrieval

Abstract

We present a cache-aided private information retrieval (PIR) problem for a replication-based storage system, where a user caches parts of messages in advance. We propose a new PIR scheme from non-colluding databases with cached data at the user. The main idea of our scheme is to exploit cached data as a side information in a specific way to minimize the amount of download for the retrieval. The amount of download of our scheme is within a multiplicative factor of 2 from the lower bound of the amount of download. Furthermore, we show that the performance gap is zero for a specific range of cache size.