Research on MP2P Resource Index Diffusion Methods
|School||Harbin Engineering University|
|Course||Computer Software and Theory|
|Keywords||MP2P network index diffusion query driven benign worm Epidemic model|
With the rapid development of the Internet and wireless communication technologiesbecoming more sophisticated, people require more and more stringent network service modeand quality. How to meet the user joining into the Internet to obtain necessary resources isone of the hot issues for the researchers. The rise of mobile computing technology, makingthe computer applications fields of use has been widely extended. People could use themobile technology to access the network at anytime and anywhere. Eliminating center server,P2P technology can satisfy the direct communication between the network nodes. Throughmixing traditional P2P technology, wireless communication technology, mobile computingtechnology together, MP2P completed seamless access to wireless network. Nodes in thenetwork can communicated with each other freely for resource sharing and collaborativeservice.However, in order to satisfy resources sharing, we first need positioning resourcesuccessfully. The existing study found that index can improve the efficiency of resourcelocation. Therefore, this paper mainly studies on the MP2P network index diffusion methods.To begin with, this paper introduces the characteristics and application fields of MP2Pnetwork and analyzes the technical problems of the existing applications.Then, it introducesthe current situation of the researches on resource positioning、index information diffusionboth at home and abroad. From two different angle-node request resources and resourcepopularity, this paper proposed passive index diffusion method based on inquires driven andactive index diffusion method based on the resource popularity.In passive diffusion method,proposed index life period mechanism to fight with index failure problem that cause bynetwork dynamic, also proposed index validation-update mechanism to prevent invalidindex diffusing in the network. In active diffusion method, we set diffusion threshold toprevent the diffusion of the index with low popularity, considering the diffusion rate andnetwork load, this paper selects benign worms-epidemic diffusion model. It can reduce theunnecessary network communication flow.At last, the paper analyzed the proposed diffusion algorithms through simulationexperiments. From the simulation results, we know in the passive diffusion algorithm, indexdensity has nothing to do with the specific search algorithm, but query arrival rate and index lifetime. In the active index diffusion algorithm, the diffusion efficiency and network load islinear relationship. Finally, we confirmed index diffusion can improve the efficiency ofresource location, achieve the desired goals.