Research Of Resource Discovery Algorithm In Mobile Peer-to-Peer Network

With the continuous optimization of mobile networks and the rapid popularization of mobile devices,people gradually begin to use mobile networks to achieve the various functions in the fixed networks.Stat- stics show that 66.2% of Internet users are mobile Internet users.Sharp rise in the num- ber of mobile users at the same time,mobile operators are also facing the problem that the speed of mobile network traffic increase is much faster than the growth rate of wire- less bandwidth.In order to effectively resolve this conflict,In order to effectively resol- ve this contradiction the major mobile operators have begun to explore the application of P2P technology in the mobile network environment which called MP2P techno- logy.But the mobile network environment has many features, such as:high dynamic, lack of node processing capacity,limited network resources and so on. So the original P2P technology can't be directly applied to mobile networks.In P2P technology the most important research direction is resources discovery,so the object of this paper is how to improve the existing search algorithm in P2P to make them adapt to the mobile network and improve the performance of resources discovery.After analysing the features of MP2P,this paper design an improved resource discovery algorihm which based on different kinds of Ant Colony and mobile agent.In addition, it also make a super node failure-handling mechanism.The nodes in mobile network have a high dynamic characteristic.When super-node become failure,the common nodes that managed by invalid super-node are forced to leave the network and the resources that stored in these nodes have been lost. To improve recall of resources,the paper proposes an alternate node mechanism.The mobile agent in this algorithm can make nodes interact with each other by a non-direct way which can reduces the bandwidth consumption of the search.Heterogeneous multi-group ant colony algorithm is another innovation of the algorithm, it divide the whole colony into multiple sub-ant colonies which can belong to different ant colony systems.The design can keep balance between the convergence speed and the diversity of result.The existing historical information can be made good use of with pheromone in the ant colony algorithm.This way can reduce search Latency and network load. At the end of paper MATLAB is used to test the validity of this improved algorithm. The results show that the improved algorithm can improve resource discovery recall rate and reduce traffic load,and moreover this advantage is more apparent with the increasing scale of the network.