Mitigating On-Off attacks in reputation-based secure data aggregation for wireless sensor networks

  • Authors:

  • Hani Alzaid;Ernest Foo;Juan González Nieto;Ejaz Ahmed

  • Affiliations:

  • Information Security Institute, Queensland University of Technology, GPO Box 2434 (126 Margaret Street), Brisbane QLD 4001, Australia;Information Security Institute, Queensland University of Technology, GPO Box 2434 (126 Margaret Street), Brisbane QLD 4001, Australia;Information Security Institute, Queensland University of Technology, GPO Box 2434 (126 Margaret Street), Brisbane QLD 4001, Australia;Information Security Institute, Queensland University of Technology, GPO Box 2434 (126 Margaret Street), Brisbane QLD 4001, Australia

  • Venue:
  • Security and Communication Networks
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

In-network aggregation is considered as an efficient way to reduce the energy consumption in wireless sensor networks (WSNs). However, it opens doors for a compromised node to distort the integrity of the aggregated data by altering the data and disrupting transmission of the aggregation results. Thus, several secure data aggregation protocols were designed to mitigate the effect of the node compromise attack and ensure data integrity. Most protocols can detect the manipulation of the aggregation results and then reject them at the base station, which gives a single node compromise the opportunity to disrupt the limited resources in the network. Reputation-based secure data aggregation protocols take a step further in helping to identify compromised nodes as early as possible. However, reputation-based protocols are prone to On-Off attacks (OOs) in which a compromised node is able to affect the aggregation results without being detected. The compromised node behaves maliciously now and then to ensure that its reputation value is within the trustable level. A solution to defeat this attack is proposed in this paper. The significance of the proposal is twofold: (i) it extends Alzaid et al.'s protocol and mitigates the effect of the OO on the aggregation results, and (ii) it considers non-homogeneous environments, which requires distinguishing between abrupt and incipient changes. In a comparative analysis of our proposal with Alzaid et al.'s protocol, plain estimate, and reputation-based estimate shows its superior performance in mitigating the effect of the attack. Copyright © 2011 John Wiley & Sons, Ltd.