부산대학교 도서관

Centralized networks can perform task scheduling readily because they can easily collect system information and manage tasks. However, due to the existence of the central nodes, centralized networks have low fault tolerance. On the contrary, peer-to-peer(P2P) networks have higher fault tolerance, but they do not have global information without central nodes. Therefore, how to schedule tasks in a P2P network has become a problem and a challenge. This paper proposes a consensus algorithm to maintain task information in the P2P network and implements a scheduling system to solve this problem. At the same time, it combines the ease of container deployment and migration to improve the efficiency of task scheduling. Besides, the heartbeat mechanism between nodes achieves fault tolerance for nodes. The system can be used in network environments without central nodes, such as mobile Ad hoc networks.