In the past few years there has been a major leap in the size of data that exist on the Web. Especially the launch of Web 2.0 websites like Wikipedia.org or YouTube.com had its effect on the presented information complexity as well as its actuality. There has been also a steady rise of B2B transactions enabled by the exchange of data between companies. The actual acceptance of e-commerce applications increased the amount of transactions in the B2C segment as well. As these changes already affected the structures of today’s Enterprise Architectures (EAs), future architectural concepts in this context will have to pick up challenges which are caused by the dynamic availability of a huge amount of services as variable elements of a next generation of enterprise software systems and system landscape. The article focuses on the problem of the dynamic discovery and invocation of appropriate Web Services in EA environments to fulfill this purpose. The proposed model is based on a decentralized network of Web Service providers and uses software agents which are responsible for the intelligent discovery and management of Web Service-Endpoints suitable to the needs of the consuming entity. The model considers preferences of users as dynamic criteria. Furthermore the reputation of service providers is seen as collective knowledge of a user community whereas past experiences of neighbored users influence the autonomous acting of user agents. Thus, relevant Web Services will be semantically described and discovered in a collaborative manner. The proposed architectural model is based on an open network architecture which abandons central elements for the management of service offerings in order to increase the openness of conventional architecture approaches. This article shows what current SWS technology offers in respect to the requirements needed by future EA scenarios and will point out the potentials of current SWS technologies in the EAI and B2B context. Furthermore SWS frameworks will be investigated according to their potential as regards to a dynamic discovery and invocation in the given context.
Journal of Enterprise Architecture