Context-Aware Service Adaptation: An Approach Based on Fuzzy Sets and Service Composition

service composition Architecture specific service composition Execution and monitoring User’ context Composition request management Fig. 5. Overview of the proposed service composition architecture. Fig. 5 above shows an overview of the system architecture, which employs a process approach to service composition, to fulfill the design requirements outlined in the previous section. The four layers in the system architecture map to the four main stages in the service composition process. The first layer is the composition request management layer, which assembles and, if necessary, modifies a composition request. Each composition request is a formal definition of the user’s task intention. The next layer is the abstract service composition layer, which generates an abstract plan. An abstract plan is a set of abstract services and their control flow, comprising the composite service. Abstract services are high-level descriptions of service operations and cannot be directly invoked. The architecture specific service composition layer instantiates the abstract plan and generates a deployable service description, which represents a service instance. The deployable service description is passed to the execution and monitoring layer, which invokes the specified service instance and monitors its execution. 6.1 Composition Request Management Layer Fig. 6 shows the structure of the first layer of the composition framework. The comCONTEXT-AWARE SERVICE ADAPTATION 11 position request is an entry point to the composition process. It specifies the user’s task and consists of two parts. The first part is a description of the core user task, for example, Mohcine’ request in Case 1 of the usage scenario for directions to the nearest Lebanese restaurant, selected from the Goal Service (step 1). The second part contains contextual parameters. For example, if Mohcine is using an android smartphone and he is currently moving near the shop center that would specify the computing device and the location, such contextual parameters further customize the composition request. For instance, in this context, it may be more appropriate to read out the driving directions to Mohcine. This layer constructs the composition request and feeds it to the abstract service composition layer (step 3). If the abstract service composition process fails (step 4 (a)) control is passed back to the composition request management layer, which attempts to transform the composition request into an alternative request that can be satisfied. For example, Mohcine’ original composition request to find the nearest Lebanese restaurant may be replaced by a more generic request of finding any type of restaurant nearby. Furthermore, the functionality to present the output in speech format may be added if the speech synthesizer service is available and Mohcine makes the kit. Composition request Context Service Goal Service Goal Transformation Engine Context Mesh Context Proxy Service recommender Abstract service composition Architecture specific service composition Execution and monitoring Composition request management 5b: transform context goal condition