Autonomous service level agreement negotiation for service composition provision

Efficient management of service level agreements which specify mutually-agreed understandings and expectations of service provision has been a subject of research for a few years. A critical issue in this area is for service consumers and service providers to effectively achieve agreements on non-functional aspects of service provision, such as quality of service. However, this issue has not been well addressed, especially in the context of service composition provision which implies the establishment of a set of interrelated agreements on quality of service between the service consumer and multiple service providers offering various services in the composition. There is a lack of supporting frameworks and techniques to automatically and dynamically achieve agreements on quality of service constraints for individual services in a service composition, aiming at fulfilling composition's end-to-end quality of service requirements. This paper reports the authors' recent research in addressing this issue, using the agent technology. In this research, the service level agreements for a service composition are established through autonomous agent negotiation. To enable this, an innovative framework is proposed in which the service consumer is represented by a set of agents who negotiate quality of service constraints with the service providers for various services in the composition. This negotiation is well coordinated in order to achieve end-to-end quality of service requirements. Based on this framework, a new negotiation protocol is presented to support coordinated negotiation. A utility-function-based decision-making model is proposed based on which agents can proactively decide on the course of further actions. Moreover, this paper also contributes the novel design of the negotiation Web service on the service providers' side for the purpose of interoperability. Finally, the prototype implementation for the purpose of proof-of-concept is discussed. (C) 2007 Elsevier B.V. All rights reserved.