Adaptive Scheduling of Message Forwarding of DHT-Based P2P Network Broadcast

摘要

Aggregated information, such as the total and free capacity of storages, computing power, network bandwidth, and so on, is very important for many peer-to-peer (P2P) applications and P2P-based grid or cloud computing. Previous studies have proposed to build a broadcast tree for a structured P2P network. Along this tree, the root can broadcast a command message to all the other peers, and in the reversed redirection, information can be aggregated from leaves to the root. Although command broadcast and information aggregation can be processed within O(log N) hops expectedly, where N is network size, the required time may vary dramatically. In this paper, the authors furthermore consider the order of message forwarding in a broadcast tree. A low-cost fully-distributed algorithm running on each peer is presented to adaptively schedule the message forwarding. The simulation result shows that the proposed algorithm can significantly reduce the time required to complete command broadcast, as well as information aggregation. In addition, the load of peers is furthermore balanced.