CPA: A Cluster-Based Packet Acknowledgement Approach in WSNs

Abstract

In wireless sensor network, repairing packet loss lead to increasing of the network load and substantial energy consumption. In order to reduce the network load, most of transport protocols employ a Cumulative Acknowledgment (CACK) policy for received packets. In this paper, we propose a Cluster-based Packet Acknowledgement (CPA) approach that naturally reduces more ACKs than traditional approaches of cumulating multi-packets of multi-nodes. A distributed greedy clustering algorithm is employed to form clusters with the minimum number. Different strategies are used to reduce the amount of packet ACKs in intra-cluster and inter-cluster respectively. And a scalable Bloom Filter is used to fix the length of the CACK, which can effectively save the storage. We also discuss a guarantee condition for the reduction of packet ACK path length. Our simulation results show that the CPA approach effectively increases the packet delivery ratio, and also reduces both the average end-to-end delay and the total length of ACKs, which thus decreases the network load.