Information-Theoretic Approach to Estimating the Capacity of Distributed Memory Systems

Systems with cash memory (or more generally, with distributed memory) are very widely used in information technologies. Such are content delivery networks (CDN) of various types, which deliver digital movies, books, and similar content; peer-to-peer (P2P) networks, where millions of members exchange various information; and many other systems and devices of this kind. We introduce the notions of capacity and entropy efficiency for distributed memory systems, propose methods for estimating these quantities, and give an example of their application.