Reputation Assessment

Abstract

In this chapter, we describe the assessment module of the RATEWeb framework: a Reputation Assessment framework for Trust Establishment among Web services. The focus is on providing a comprehensive solution for assessing the reputation of service providers in an accurate, reliable, and decentralized manner. Since reputation forms an integral part of service-oriented environments in relation to the dynamic selection of services, on-the-fly composition of value-added enterprises, and optimization of service tasks, we have chosen Web services as a representative domain. However, RATEWeb can be extended and used in other contexts and domains. The proposed framework takes into account the presence of malicious raters that may exhibit oscillating honest and dishonest behaviors. Previous solutions for reputation assessment make simplifying assumptions that may not apply in a service-oriented environment. For example, [66] relies on pre-existing trusted parties, in [31] and [20] data needs to be distributed according to a certain statistical distribution, a common set of past providers is required in [146] for evaluating rater credibility, and in [95] human intervention is required, meaning the assessment process is not fully automated. Other similar solutions either do not consider all facets of reputation [1, 65, 116] or are focused primarily on efficiency/performance (rather than functionality) [36]. We develop a simple and holistic solution that provides an automated and adaptive reputation mechanism, whereby reputations are evaluated through a number of heuristics with different perspectives providing a fair and accurate assessment [86].