Research On Task Scheduling Strategy Based On Rough Set Theory In Peer-to-Peer Environment

Each node of the P2P network is both a resources provider and a resources consumer. All of nodes in the P2P network can freely join or leave the network, with strong autonomy and randomness. Thus the core of the whole Internet's network application has transferred from central server to terminal device on net edge. People can take part in net activities more initiatively, which realizes the freedom and equality principle of the Internet. Meanwhile the utilization of sources like information and band width is improved. Task scheduling is a key technology in P2P computing, it is a direct impact on the computing performance of the entire system.In P2P environment, it is difficult to obtain complete information of task and node. More than one property is needed to describe the task and node, and there may be irrelevant attributes exist. In this paper we first use rough sets to extract feature vectors from the attribute of the tasks and nodes, on the basis of which determine the type and assign theirs category which with the largest similarity with it, according to the similarity formula. In the P2P environment, it should first select the spare node sets that meet the requirements of task scheduling time before the task being scheduled. However, whether the node is idle or not at a particular moment can not serve as the condition for task scheduling, the idle time interval of the node is also should be considered. In this paper, a statistical method is used to obtain the node's idle time interval. On the basis of ensuring the successful implementation of the task, which has facilitated improved scheduling efficiency. In addition, select the nodes that are suitable for task scheduling depend on time by adjusting confidence coefficient dynamically.After put the division of tasks and nodes in different queues and collections, this paper defines the task scheduling module based on queue theory, and sort the task of the same queue according to dynamic priority. According to task scheduling mechanism for selecting spare nodes, a task queue dispathc to the same set of nodes, so completed task scheduling. Finally, take the average QoS evaluation index as task scheduling, which mainly includes the task execution time, communication time and cost. Experimental results show that the task to improve the average QoS in this paper.