Analyzing Communities of Web Services Using Incentives

Introduction

As one of the recent technologies for developing loosely-coupled, cross-enterprize business processes (usually referred to as B2B applications), a plethora of web services exists on the web waiting to receive users' requests for processing. Such requests are usually competitive in a a security and reputation-driven environment (Martino & Bertino, 2009; Zhang, 2008). To this end, the reputation assessment has been addressed in recent proposals (Jurca & Faltings, 2003; Jurca & Faltings, 2007; Liu et al., 2004). One general solution for such reputation assessment is collection of the after-interaction feedback that users provide with respect to the quality of the received service. However, in feedback-based reputation mechanisms, the precise reputation assessment needs to be verified. Selfish web services might manage to provide feedbacks that support them in the reputation mechanism. In general, online reputation mechanism is always subject to get violated with selfish web services. Another way to address the selection (and management) problem is to gather web services having similar functionalities to a community. Community of web services (CWSs) is a gathering of single and functionally similar web services that are aggregated to perform as one community while offering unique or variety services. The main property of a CWS is to facilitate and improve the process of web service discovery and selection and effectively regulate the process of user requests. There are underlying reasons for this. In general, the individual web services fail to accept all the requests for them, and thus refuse to accept a portion of their concurrent requests. This would decrease their overall reputation in the environment and would lead to loose some users. In CWSs, the community gathers a set of functionally homogeneous web services. Given that some communities offer the same functionality (hotels booking, weather forecasting, etc.), there is a competition between different communities. In this case, reputation is considered as a differentiation driver of the communities. Moreover, reputation helps users to select the most reputable community, which would provide the best QoS, and helps providers to join the best community, which would bring them the most value. Users assess the reputation of the community and upon that request for a service. Although the service selection process might be simplified, still communities might distract the reputation mechanism to support themselves. To this end, the reputation mechanism is needed to maintain a truthful service selection procedure.