Abstract
Service composition is the process of constructing new services by combining several existing ones. It considered as one of the complex challenges in distributed and dynamic environment. The composition process includes, in general, the searching for existing services in a specific domain, and selecting the appropriate service, then coordinating composition flow and invoking services. Over the past years, the problem of web service composition has been studied intensively by researchers. Therefore, a significant amount of solutions and new methods to tackle this problem are presented. In this paper, our objective is to investigate algorithms and methodologies to provide a classification of existing methods in each composition phase. Moreover, we aim at conducting a comparative study to discover the main features and limitation in each phase in order to assist future research in this area.
Highlights
Service-oriented architecture (SOA) is a structural model composed mainly of services
In a Multi-Agent System (MAS), service composition is achieved by the coordination between multiple agents to discover, build, select, and execute composite service considering semantic of services parameters
Workflow-based planning Two main methods used in composition planning as workflow: Workflow net and Business process modeling notation (BPMN)
Summary
Service-oriented architecture (SOA) is a structural model composed mainly of services. The phase, the service discovery, a search for services that match the functionality and non-functionality requirements for each task in the composition is performed. From the multiple services discovered in the previous phase, service selection comes to select the most appropriate service for each task in the composition in order to satisfy user requirement. We will present a broad taxonomy of service composition methods and phases by survey and classify the existing solutions for each composition phase. This will allow us to identify the different approaches and discuss their adaptability and capability to solve such a problem.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have