Deadlock Detection in Distributed Database

Volume 2, Issue 2 March – April 2013 Page 95 Abstract: In distributed database, data resides in various sites and many transactions can originate at any number of sites randomly. These transactions can execute concurrently. This concurrency leads to deadlock in which transactions may enter into an infinite waiting state so deadlock handling is an important criteria in distributed transaction processing. So, an efficient algorithm for detecting deadlock is to be devised. Various approaches are there for detecting deadlock in distributed database such as chandy & Mishra Algorithm [5], Sinha’s Algorithm [8], Obermack’s Algorithm [3].All of which have been implemented and tested in distributed database where data is distributed among various sites and data is not replicated. As data replication improves availability, it is necessary to implement and test the performance of deadlock detection algorithm in a replicated environment.