Research On Key Technologies Of P2P Energy Saving

In today's Internet,P2P as a content distribution and sharing business occupies a significant proportion of traffic,and energy consumption issues have become a hot issue.In the past 20 years,a large number of users have uploaded and downloaded files through P2P services every day,accounting for 40%-73%of the total Internet traffic.In the Internet where file distribution and shared service traffic dominates,the issue of energy consumption has gradually become an important obstacle to the rapid development of the Internet.This topic analyzes the more popular research,most of which focus on energy conservation and neglect reliability.Therefore,it is necessary to optimize both the performance indicators of the network and the energy saving.We have established the ACTE-FD(Adjustable and Controllable Traffic Engineering File Distribution)by adding the adjustable factor.The model regulates the energy-saving and reliability of the network by concentrating or even spreading the traffic in the network.Based on the model,this paper presents a set of parallel computing methods based on dual decomposition and sub-gradient projection.It can solve the network nodes of a certain scale quickly,but the solution time for 2000 nodes will reach the hour level.Therefore,on the basis of this algorithm,the GPU is introduced to speed up the calculation and deflect the sub-gradient projection to reduce the number of iterations,even if the problem scale reaches nearly 10,000 nodes,it can be solved faster.Finally,the solution distribution and the specific network topology are given,and a set of deployment scheme based on BT protocol is given.The experimental design of ACTE-FD was performed to verify the performance of the model.The results show that in terms of energy saving and reliability,by adjusting the adjustable factor,the network energy consumption is reduced by 47.09%in the pursuit of network energy saving.Compared with similar foreign research ECFD,the power consumption is about 10.8%lower.In the pursuit of network reliability,the reliability of the network increased by 58.18%.After the introduction of GPU degree,the computational efficiency is significantly improved,and the computation time of the parallel solution method is reduced by 95%.After introducing the deflection sub-gradient algorithm,the algorithm time is reduced by nearly 50%compared to the standard sub-gradient algorithm.