Distributed Reasoning Research Based On Named Data Network

With the development of network and communication technology, the Internet has evolved into a ubiquitous information infrastructure for content distribution and service providing. However, intelligently discovering the knowledge and rules is a challenging problem in the large-scale distributed and dynamic data sources in the Internet. The network protocol system, which is data oriented and centered on the content or information, realize the change from the "machine interconnected" to "information interconnected". In this thesis, we mainly studied the distribution, self-organizing semantic reasoning method, algorithm and system implementation based on the named data network architecture and the large position-independent data namespace. The main research works were as follows:First, for applying on the named data network environment, we researched the user-driven semantic reasoning mechanism. In order to realize the name-based routing on the dynamic, distributed network environment, the named data network separated the name and the location of the content. In this point,(1) for finding all data sources that relevant to a user in dynamic distributed network environment, we researched the semantic forwarding process of the user’s reasoning request, and designed the forwarding engine with reasoning ability in the forwarding nodes;(2) for backing and convergencing the distributed reasoning results, we researched the semantic integration and convergence mechanism and implementation method;(3) due to the reasoning process are needed in the distributed, self-organizing environment, we also studied the reasoning mechanism and algorithm with dynamic adaption and high efficiencySecond, we researched the demand-driven reasoning rules for semantic reasoning during routing. Semantic reasoning should be simple and effective in the dynamic, distributed network environment. Firstly, under distributed environment, according to the reasoning demand, we proposed the expression method of reasoning rule that is driven by request; Then, based on the hierarchical naming mechanism and the features of the route index, we put forwarded the request’s forwarding reasoning mechanism, for discovering all potential sources of knowledge and forwarding the request to them; Finally, based on the results aggregated at the convergence point, we researched the semantic integration rules for refining and integrating the reasoning results, and studied the evolution rules of the reasoning request, for finding more potential knowledge the user needed according to the acquired knowledge.Third, we researched the implementation method of the dynamic, distributed semantic forwarding and integration system based on the named data network. In the dynamic, distributed network environment, semantic forwarding and integration require multiple nodes to cooperation and coordination. First of all, we simulated multiple forwarding nodes and data sources. And then we realized the self-organizing semantic forwarding and results integration in the actual network. Finally, we evaluated the efficiency of the forwarding and integration and compared our system and the current works.The research of this thesis will solve the dynamic, distributed reasoning and the issues of position-dependent, effectiveness and accuracy on knowledge discovery, the issues of the effectiveness on content delivery in the network environment. Based on the technology and algorithm presented by this thesis, we proposed a distributed reasoning framework, process and realization model based on named data network, solved the discovery and integration problem of the dynamic, distributed knowledge, and provided the better theory and technology basis for the large-scale knowledge discovery, use and management under the network environment.