A Redundancy Protocol for Service-Oriented Architectures

Abstract

Achieving high-availability in service-oriented systems is a challenge due to the distributed nature of the architecture. Redundancy, using replicated services, is a common software strategy for improving the availability of services. However, traditional replication strategies are not appropriate for service-oriented systems, where diverse services may be grouped together to provide redundancy. In this paper we describe the requirements for a redundancy protocol and propose a set of processes to manage redundant service providers.

Dynamic service quality and resource negotiation for high-availability service-oriented systems

“…We have modeled two primitive service redundancy strategies and showed how they can be composed into complex, hierarchical redundancy groups [12]. In addition, we have specified a protocol for managing redundant services in an SOA [11]. Finally, we have implemented a method to generate RADL models from SCA specifications.…”

Approaches to Improve Reliability of Service Composition

“…If data is modified it is important that changes are performed only once. This implies the need for compensation, transactional behaviour or two phases commit (see [12]). The call is successful if at least one service provider is performing well.…”