Research Of Cache Management Mechanism And Traffic Engineering In Information-Centric Network

ICN(Information-Centric Network) is a new network architecture considering “content” as the “ firsr class citizen” insteading of “IP address”, and NDN(Named-Data Network, NDN) is a typical implementation of ICN. This thesis studies the cache management and traffic engineering problem in NDN.PIT(Pending Interest Table), which is one of the caches in a NDN router, is used to record the information of Interests that are unsatified. Moreover, PIT can aggregate the duplicated Interest packets and provide native support for multicast service, which can reduce network bandwidth consumption. However, due to the line-speed processing requirement of PIT in real implementation, the PIT capacity in practical routers is very limited, but, in fact, PIT storage space required is enormous due to the largre number of request packets. Therefore, how to reduce the required capacity of PIT, while ensuring getting good performance at the same time has become an important issue. We observe that PIT can get good performance by only store popular requests according to the theratical analysis of the hit ratio in idealized PIT, which whose capacity is umlimited. And based on this observation, we proposes a PIT management strategy based on content popularity. The proposed strategy uses hybrid forwarding mechanisms to forward Interests according to their popularity, such that PIT can keep a high hit ratio with small size. The simulation verify that the strategy is feasible and has promsing forwarding performance.CS(Content Store) is the key to improve the efficiency of content distribution in NDN. NDN uses LCE(Leave Copy Everywhere) caching strategy, which may lead to a large amount of caching redundancy and has unstatified caching performance. Inspired by the stydy of chapter 2, we proposes a CS management strategy. The proposed cache management strategy caches different content at different routers according to its popularity, which can have good performance and avoid the redundant contents problem at the same time. In addition, this thesis also proposes a comprehensive cache management strategy putting both PIT and CS into account. And the simulation verify that both strategies are feasible and have promsing forwarding performance.For traffic engineering problem, we proposes a NDN load balancing scheme basing on multi-topology routing. The proposed scheme balances the load through protecting the link which has heavy link load by routing strategy and CS. And the simulation verify that the strategy is feasible and has promsing performance by testing different multi-topology generation scheme.