Distributed in-network cooperative caching

摘要

Named Data Networking (NDN) featuring in-network caching capability is a large effort that exemplifies information-centric approach to networking by shifting emphasis from hosts to data so as to meet growing demand on content. This work explores a scheme that enables a NDN domain to make full use of its in-network caches to enhance its performance, availability, and reliability. Currently, each NDN router independently determines what contents to cache and is unaware of content cached in nearby routers and thus their caches are not utilized in an efficient way. This paper proposes to have routers in a NDN domain share cached data and coordinate to make caching decisions (which is dubbed in-network cooperative caching) and formulates it into a constrained optimization problem. The Lagrangian relaxation and primal-dual decomposition method is applied to decompose the optimization problem into object placement subproblems and object locating subproblems, each of which can be solved in a distributed manner at each router, such that the in-network cooperative caching is addressed in a distributed way. Our simulation results, although preliminary, suggest that our scheme can benefit users, Internet Service Providers (ISPs) as well as content servers, and the improvement can be as much as 88% compared to current NDN caching policy.