Efficient clustering approach for intrusion detection in adhoc networks

Abstract

Mobile ad-hoc networks (MANETs) are temporary
wireless networks useful in emergency rescue services,
battlefields operations, mobile conferencing and a variety of
other useful applications. Due to dynamic nature and lack of
centralized monitoring points, these networks are highly
vulnerable to attacks. Intrusion detection systems (IDS)
provide audit and monitoring capabilities that offer the local
security to a node and help to perceive the specific trust level
of other nodes. We take benefit of the clustering concept in
MANETs for the effective communication between nodes,
where each cluster involves a number of member nodes and is
managed by a cluster-head. It can be taken as an advantage
in these battery and memory constrained networks for the
purpose of intrusion detection, by separating tasks for the
head and member nodes, at the same time providing
opportunity for launching collaborative detection approach.
The clustering schemes are generally used for the routing
purposes to enhance the route efficiency. However, the effect of change of a cluster tends to change the route; thus degrades the performance. This paper presents a low overhead clustering algorithm for the benefit of detecting
intrusion rather than efficient routing. It also discusses the intrusion detection techniques with the help of this simplified clustering scheme.