Managing failures in web services atomic commitment protocols

WS‐ReliableMessaging specification describes a protocol that allows messages to be delivered reliably between distributed applications in the presence of software component, system, or network failures. However, it ensures reliable communication only in the context of two sites – it does not provide any means for consistent termination of the executions spanning over more than two sites. This paper aims to address this issue.

The paper presents the Reliable WS‐AtomicTransaction protocol, and illustrates its implementation by exploiting WS‐Coordination, which describes an extensible framework for providing protocols that coordinate the actions of distributed applications. The paper also presents the ontology of the log, which is maintained by the Reliable WS‐AtomicTransaction protocol. The ontology is presented in a graphical form and in OWL.

The introduction of an atomic commitment protocol and its termination protocol increase the reliability of the executions of distributed applications in service‐oriented architectures. On the other hand, it complicates the management of distributed applications as the atomic commitment protocol has to maintain the log that is used by its termination protocol.

The paper presents an atomic commitment protocol and its termination protocol, which is failure resilient and non‐blocking as long as a failed site can communicate with a process that has received sufficient information to know whether the transaction will be committed or aborted. Decreasing the amount of blockings is important because blocking can cause processes to wait for an arbitrarily long period of time.