Research On Label-directed Forwarding And Cache Scheduling In ICN

With regard to design of future network architecture, the research area Information Centric Network (ICN) attracts a great many researchers with the advantage of better meeting the user’s needs that most users prefer to obtain the contents of information rather than the location of information. As the most important and promising technology realizing ICN, Named Data Networking (NDN) can provide technical support for multi-user content dissemination beyond what IP architecture supports. However, current name-directed routing and forwarding process of NDN does not satisfy the transmission efficiency requirements of users; meanwhile, the high-speed in-network caching resources are not utilized smartly enough, although in-network caching schemes conceptually reduces the average transmission distance for content retrieval. Our work is just developed based on this background.This thesis propose solutions for optimizing the forwarding and caching schemes of NDN which effectively optimize the processes of Interest message forwarding, data-path establishing, and caching status monitoring; it also optimizes the utilization efficiency of in-network cache resources. The schemes proposed in this thesis better satisfy the applications that require high bandwidth and low transmission latency, especially real-time multi-user applications in ICN.Our major work includes:First, we design a centralized routing calculation system and build the control channel based on the "Interest-Content" communication model of NDN on the basis of analyzing the routing management system of ICN. Implementation of our design on ndnSIM proves that our system can effectively reduce the cost of data-path establishment and synchronization of caching information in current centralized routing architectures.Second, a novel scheme is designed in which interest messages are forwarded according to label sequences with variable lengths on the basis of analyzing Interest message forwarding process on NDN nodes. Implementation of this scheme and numerical experiments for efficiency evaluation have been conducted on ndnSIM.Last but not least, based on analyzing the features and defects of related research on in-network caching and theoretical reasoning, we propose a cache scheduling scheme, Cache When Necessary (CWN). According to the simulation results, our CWN scheme efficiently reduces the redundancy in cache resource consumption without introducing higher average transmission delay in content retrieval.