Examination Committee

Prof Danny H K TSANG, ECE/HKUST (Chairperson)
Prof Chin-Tau LEA, ECE/HKUST (Thesis Supervisor)
Prof Ling SHI, ECE/HKUST

Abstract

Content Centric Networking (CCN) is a new network paradigm which shifts the Internet’s central abstraction from the IP to the named content. One of the most important features of CCN is the ubiquitous in-network caching: every node in CCN has the capability to cache the content passed through it. This feature brings lots of advantages to CCN such as delay and traffic load reduction, but it also brings challenges to the routing design. In this thesis, we study the optimal routing problem for three different scenarios in CCN.

The first scenario we study is the single-server CCN. We propose an analytical model to minimize the congestion ratio under single-path shortest path routing. We come up with an approach to approximate the cache hit ratio and the link load based on the shortest path routing tree. Then we employ a heuristic scheme called simulated annealing to find the nearly optimal link weight setting. The simulation result shows that our approach can significantly reduce the congestion ratio compared to the conventional routing scheme in an IP network.

The second scenario we examine is the multiple-server CCN. This time the contents are distributed among all nodes in the network. Each node has its own storing content and the capability to cache the content from other nodes. An analytical model based on multi-path shortest path routing and a heuristic approach are proposed to minimize the congestion ratio given the traffic matrix.

Finally we study the optimal routing in CCN under the hose traffic model. In this case, the exact traffic matrix is unknown. The only constraints are the maximal amount of ingress and egress traffic at the edge nodes. We design a non-blocking content-centric network which maximizes the throughput within the edge traffic constraints.