Service Discovery for Service-Oriented Content Adaptation

Service-Oriented Content Adaptation (SOCA) has emerged as a potential solution to the content-device mismatch problem. One of the key problems with the SOCA scheme is that a content adaptation task can potentially be performed by multiple services. In this paper, we propose an approach to the service discovery problem for SOCA and it is demonstrated to perform well. 1 Background and Related Work Content adaptation is a multi-step process involving a number of services each performing a specific adaptation task. The service discovery problem of interest can be formulated as follow: Let , , ... , and , , ... , be a set of adaptation tasks and a set of available services respectively such that . Let Q represent the quality of service criteria (e.g., time, rating, reputation, cost) of each available service . Given S, T and Q, the central problem is how to discover and select a set of composite services that are capable of performing series of tasks. Existing QoS-based service discovery systems for web service [1] and for pervasive computing require that both the services and the client QoS must be known a priori. However, this may not be practical as most end-users may find it difficult to indicate all their QoS precisely. Moreover, there may not be services that match users’ indicated QoS thus, resulting faulty QoS matchmaking. The proposed approach addresses these problems. 2 Proposed Solution and Evaluation Fig.1(a) shows a high level service discovery architecture for the SOCA platform. The provider layer has the platform providing the services. A service provider advertises its services in one or more service registries. Access to a service is made possible through its service handle in a reference form. The discovery layer deals with those aspects that describe how service discovery is operated to facilitate incoming client requests. Fig.1(b) outlines the algorithm for the service discovery. It is composed of three inter-related steps: adaptation function matching, candidates’ discovery and composite service selection. For details of the proposed discovery architecture and algorithm, please refer to [2]. * Faculty of Comp.Sc.& Info.Tech., Tun Hussein Onn University of Malaysia. Service Discovery for Service-Oriented Content Adaptation 309 Clients Service Discovery List L