A highly-reliable data recovery model based on parallel transmission for a P2P distributed system

The data failure model given by interleaving memory and parallel transmission can achieve the goal of reliable storage and fast data recovery, but the assumption that the download speed increases linearly with the increase of the number of parallel transmissions is too ideal. To overcome the problems from past research, this paper proposed a data recovery model based on the parallel transmission strategy, Poisson distribution theory, and the mechanism of a data exchange center. The model introduced the non-linear speed and function, cancelling the constraints and assumption for parallel transmission speed of a physical network. The paper theoretically analyzed the influence of all parameters on the performance of the data recovery model. The experimental result is consistent with the theoretical analysis. On the premise that the storage space used by the model is the same as a double copy, and that the influence of network flow is identical, the model effectively reduced the data failure probability.