An election algorithm to ensure the high availability of leader in large mobile ad hoc networks. Issue 2 (4th March 2018)

• Record Type: Journal Article
• Authors: Sharma, Shantanu; Singh, Awadhesh Kumar
• Is Part Of: International journal of parallel, emergent and distributed systems. Volume 33:Issue 2(2018)
• Page Start: 172
• Page End: 196
• Publication Date: 2018-03-04
• Subjects: Distributed computing -- fault tolerance -- high availability -- leader election -- mobile ad hoc networks; Parallel computers -- Periodicals; Electronic data processing -- Distributed processing -- Periodicals; Computer algorithms -- Periodicals; 004.35
• DOI: 10.1080/17445760.2016.1191077
• Languages: English
• ISSNs: 1744-5760

Abstract

Abstract: A fundamental problem of distributed systems, leader election, is presented in the context of mobile ad hoc networks (MANETs). In many distributed systems, the presence of a leader is necessary in order to monitor underlying computations, guarantee quality functioning, take checkpoints, generate the lost token, detect quiescence conditions, etc. Hence, several leader election algorithms have been proposed in the literature. Although, most of the algorithms focus on reducing the control message (messages that have the highest priority to deliver) count, there have been almost no attention on ensuring high availability of a leader despite various types of failures, especially, in the scenarios like rescue and warfare, where the absence of the leader, even for a short duration, may lead to havoc. We focus on this issue, particularly, for large MANETs, where a large number of applications fails to perform in the absence of a leader. We present a leader election algorithm for large MANETs. The algorithm is inspired by the concept of prevailing parliamentary democracy and elects three best-nodes – in terms of performance parameters like battery life, computing power, memory, hop distance, and mobility – as the president, leader, and vice leader. The president node works as the leader of the network, in case, the leader and the vice leader both become unavailable simultaneously. On the other hand, the leader node serves all the requests. Further, we create a house of elite nodes, which ensures the presence of an executive, i.e. a leader during re-election to restrict the message overhead as well as the election latency while executing coordination related activities. Graphical Abstract: Failure of the special nodes.