Research On Key Techniques Of Server Selection And Routing In Content Centric Networks

The current Internet needs a revolutionary change because it has defect fundamentally, so the new content centric network architecture was put forward in recent years. Due to the rapid growth of the Internet content flow and user’s Quality of service requirements, content centric network is considered to be the next generation Internet development trend. The PURSUIT network system has been introduced in this paper. The PURSUIT is Publish-Subscribe Internet Technologies. On the basis of network platform the server selection algorithm and route planning algorithm is studied. Research purpose is to compare the performance of the algorithms under different conditions in network simulation, analyze the results, and to get valuable conclusion for guiding future research. This paper expounds the PURSUIT architecture, and BF forward mechanism, that implements the line speed interconnection.Based on study of PURSUIT in OMNET++ simulation software, server selection algorithms and route planning algorithms are essentially routing algorithms. They are all realized in core node in the system. But purpose is different, the server selection algorithm find destination and path between the source and destination given source point, while route planning algorithm find path between source and destination given source and destination. This leads to their very different implementation. The difference between content router and IP router is that CR really has the function of caching data, so the combination of the routing and caching in PURSUIT is particularly important. But current research on the key strategies of content centric network is still rare.For server selection, this paper discusses nearest cache point strategy, maximal duplication strategy, and minimal link utilization strategy. These strategies are all based on shortest path algorithm. The essence of two former ways is use of cache, while the third strategy lies in load balancing. High content shooting near the client side will make performance of nearest cache strategy good, but overall, maximal duplication strategy has the best shooting. Minimal link utilization strategy is the best when load is heavier and poor when the load is lighter. In general the performance of algorithms that uses cache better is also better. In addition, this paper compares three different strategies of load balance, and obtains valid conclusions.For routing planning, this paper proposes nearest cache point strategy, more cache time point strategy, more cache points strategy and rerouting strategy on cache advertisement. These strategies are not necessarily based on shortest path algorithm. They are looking for a qualified simple path to avoid the loop making BF mechanism effectively. Nearest cache point strategy is more effective when shooting high, and at the time of low shooting more cache time point strategy performs better. The performance of rerouting strategy in searching a jump even more jumps is best. More cache point strategy has the highest shooting in these strategies. Because more cache time point strategy and more cache points strategy only obtain approximate solution, this paper compares performance of the optimal solution and the approximate solution that they produce.In addition, because some algorithms of routing planning can be applied to server selection, this paper compares corresponding algorithms, and obtains valid conclusions.