Performance analysis of CCN on arbitrary network topology

Abstract

In this paper, by utilizing the MCA (Multi-Cache Approximation) algorithm, which is an approximate algorithm for numerically solving cache hit rates in a multi-cache network, we analytically obtain performance metrics of CCN (Content-Centric Networking). Specifically, our analytic model consists of multiple routers, multiple repositories (e.g., storage servers), and multiple entities (e.g., clients). We obtain three performance metrics - content delivery delay (i.e., the average time required for an entity to retrieve a content through its neighbor router), throughput (i.e., the number of contents delivered from an entity per a unit time), and availability (i.e., the probability that an entity can successfully retrieve a content from the network). Through several numerical examples, we investigate how the network topology affects the performance of CCN. Our findings include that the closer an entity is to the requesting repository, the more beneficial the contents caching is in terms of content delivery time and availability, and that the farther an entity is from the repository, the more beneficial the content caching is in terms of throughput.