iSocialMash: Convergence of social networks and services composition on a mashup framework

Facilitated by advanced Web technologies, service mashups are currently popular for composing new value-added applications. Mashup developers might often struggle to locate the relevant and appropriate services to satisfy their dynamic and personalized requirements. This paper proposes the concept of iSocialMash, a framework that assists the rapid, on-demand and intuitive composition of service mashups, by leveraging social networks. There two key observations guiding the development of iSocialMash. On the one hand, social networks can capture both common interests and personal preferences of different users in the same or similar application contexts; they might use similar candidate services and glue them together in similar manner. On the other hand, social networks can identify potential value-added composition of mashups; different users might have complementary collaboration opportunities for newly emergent goals. iSocialMash exploits the successful experiences of social networks to provide users with useful composition recommendations (such as missing or potentially proper services as well as connections between them). Capturing such composition knowledge, the users are presented by a set of ranked recommendations from which they can choose for their personalized needs. In iSocialMash, the data model leverages the social tagging to enrich the semantics and simplify the presentation and extraction of composition knowledge. We model the composition knowledge into some mashup patterns to accommodate different social networks interactions, and generate on-the-fly recommendations according to user personal requirements. We also experimentally evaluate the efficiency of our approach and present the current status of the prototype.