Modeling And Optimization For The Repair Strategy Of The Fixed Nodes In A Hybrid P2P Network

With the spread of the mobile network and growth of the mobile business, P2Ptechnology in traditional network has been extended to mobile one, as a result, a hybridP2P network including fixed and mobile nodes has emerged. Node management and noderepair have been proved to be the effective means of guaranteeing the availability,scalability and reliability of the hybrid P2P network. Considering the online/offlineprocess of the mobile nodes, based on the working principle of the repair strategy of thefixed nodes in a hybrid P2P network, continuous time queuing models are constructed andsystem performance is optimized.Firstly, the working principle of the two types of repair strategies is analyzed, and thescope of application is studied. In order to evaluate the system performance of the fixednodes repair strategy in a hybrid P2P network comprehensively, the performance measurescomposed of system failure ratio, average data access time and node switching ratio isconstructed.Sccondly, based on the hybrid P2P network, considering the online/offline process ofthe mobile nodes, the system models for describing the fixed nodes repair strategy arebuilt, and two possible states of a fixed node, i.e., online state and repairing state, aredefined. Aiming at the early repair strategy of the fixed nodes, combining the numbers ofusers to request data, online mobile nodes and online fixed nodes, a three-dimensionalcontinuous-time Markov chain is established. Aiming at the working principle of the lazyrepair strategy of the fixed nodes, taking into account the number of users request data,online mobile nodes, online fixed nodes, and fixed nodes in the repair process, a four-dimensional continuous-time Markov chain is constructed.Thirdly, by using the matrix geometrix solution method, the system models areanalyzed in steady state, expressions of the system performances, such as system failureratio, average data access time, and node switching ratio, are given. Numericalexperiments and simulation experiments are provided to illustrate the influence of thenumber of fixed nodes and the node repair threshold on the system performance. Finally, considering the trade-off between different system performances, costfunctions are built. For the eager repair strategy of the fixed nodes, the number of fixednodes is optimized, while for the number of fixed nodes and the node repair threshold areoptimally configured with the lazy repair strategy.