A deadlock detection algorithm using gossip in cloud computing environments

Abstract

To detection deadlock in distributed systems, the initiator should construct a global wait-for graph in an efficient way. In this paper, we present a deadlock detection algorithm using gossip for cloud computing environments where each node may leave and join at any time. Due to its inherit properties of a gossip protocol, we claim that our proposed deadlock detection algorithm is scalable and fault-tolerant. The amortized message complexity of our proposed algorithm is O(n), where n is the number of nodes. Our evaluation over scalable settings shows that our approach has a significant merit to solve scalability and fault-tolerance problems over existing algorithms. © 2013 Springer Science+Business Media.