Replica Consistency Maintenance Strategy Based On Ant Colony In P2P Network

With the continuous advance of science and technology, the computer network is being developed rapidly. The traditional client/server model has been unable to satisfy the needs of high growth of the network. Servers are facing lots of load pressure, often have problems such as congestion, collapse, and so on. How to reduce the load pressure of servers and the bandwidth consumption has become urgent problems to solve. As a new network, P2P has good scalability and little load pressure. It is being more and more noticed by researchers.File storage, one of the main applications of P2P networks, is also an important driving force of the development of P2P network technology. File storage applications based on P2P have become popular applications on the Internet, and have a tremendous influence on traffic and use of the Internet. As an effective measure to improve the resource availability and user access efficiency in P2P network, file replication becomes an indispensable important part of P2P file storage, and how to maintain existed replicas’consistency becomes an urgent problem needs to be addressed.As there are many problems such as slow update speed, redundancy and difficulty to repair disconnected path in present P2P replica consistency maintenance strategies, this paper summarizes the previous researches and puts forward a new replica consistency maintenance strategy based on ant colony algorithm. Using the ant colony algorithm in the natural ecology, the concept of pheromone is introduced. The behaviors and states of the ants are redefined according to the need of replica consistency maintenance. We model the process of replica consistency maintenance, define the structure of updated ants, propose the spread and update algorithm, and give the pheromone update formulas. These accelerate the update speed of replicas, reduce update redundant messages, and repair the broken update paths by making full use of the advantage of ants in finding path. For the cases of nodes churn ants encounter in the process of walk on the actual path, this article classifies the route path and gives a detailed analysis about every path which ants encounter in the update process. By simulating three key data, the average amount of update messages, update time and success update rate, the simulation results indicate that the proposed strategy can effectively reduce the update cost and the negative effect on consistency maintenance brought by nodes churn on the premise of keeping high success update rate.