A survey of fuzzy service matching approaches in the context of on-the-fly computing

In the last decades, development turned from monolithic software products towards more flexible software components that can be provided on world-wide markets in form of services. Customers request such services or compositions of several services. However, in many cases, discovering the best services to address a given request is a tough challenge and requires expressive, gradual matching results, considering different aspects of a service description, e.g., inputs/ouputs, protocols, or quality properties. Furthermore, in situations in which no service exactly satisfies the request, approximate matching which can deal with a certain amount of fuzziness becomes necessary. There is a wealth of service matching approaches, but it is not clear whether there is a comprehensive, fuzzy matching approach which addresses all these challenges. Although there are a few service matching surveys, none of them is able to answer this question. In this paper, we perform a systematic literature survey of 35 (out of 504) service matching approaches which consider fuzzy matching. Based on this survey, we propose a classification, discuss how different matching approaches can be combined into a comprehensive matching method, and identify future research challenges.