Secure Differential Data Aggregation for Wireless Sensor Networks

Data aggregation protocols are implemented in resource constrained wireless sensor networks to eliminate redundant data transmission for improving bandwidth utilization and energyefficiency of sensor nodes. This paper presents a secure differential data aggregation (SDDA) protocol for wireless cluster-based sensor networks. SDDA prevents redundant data transmission from sensor nodes to cluster-heads, whereas conventional data aggregation techniques eliminate redundant data at cluster-heads. SDDA first implements a sleep protocol which puts all but one sensor node that sense the same data to sleep, and then awake sensor nodes generate pattern codes to represent the main characteristics of sensed data. Cluster-heads first determine distinct pattern codes and then request only one sensor node for each distinct pattern code to transmit data to base station. Hence, cluster-heads perform data aggregation based on pattern codes. While sensors transmit their raw data to cluster-heads in conventional data aggregation algorithms, in SDDA, raw data sensed by sensor nodes are compared with reference data values and then only the difference data are transmitted. Reference data are taken as the average value of a number of previous readings. The fact that SDDA does not require cluster-heads to examine the sensed data for data aggregation helps implementing secure data transmission because the encrypted data of sensor nodes need not be decrypted by cluster-heads along the transmission path. By taking advantage of deployment estimation of sensor nodes, SDDA establishes secure connectivity among sensor nodes without performing any online key distribution. Performance of SDDA is evaluated in terms of bandwidth efficiency and energy consumption. Simulation results show that SDDA is more energy and bandwidth efficient than traditional data aggregation techniques, and the efficiency of SDDA increases as the data redundancy increases. Index Terms – Wireless sensor network, data aggregation, security, energy efficiency, base station, coverage.