Web Services Conversation Adaptation Using Conditional Substitution Semantics of Application Domain Concepts

Internet of Services (IoS) vision allows users to allocate and consume different web services on the fly without any prior knowledge regarding the chosen services. Such chosen services should automatically interact with one another in a transparent manner to accomplish the required users' goals. As services are chosen on the fly, service conversations are not necessarily compatible due to incompatibilities between services signatures and/or conversation protocols, creating obstacles for realizing the IoS vision. One approach for overcoming this problem is to use conversation adapters. However, such conversion adapters must be automatically created on the fly as chosen services are only known at run time. Existing approaches for automatic adapter generation are syntactic and very limited; hence they cannot be adopted in such dynamic environments. To overcome such limitation, this paper proposes a novel approach for automatic adapter generation that uses conditional substitution semantics between application domain concepts and operations to automatically generate the adapter conversion functions. Such conditional substitution semantics are captured using a concepts substitutability enhanced graph required to be part of application domain ontologies. Experiments results show that the proposed approach provides more accurate conversation adaptation results when compared against existing syntactic adapter generation approaches. 1. Introduction Internet of Services (IoS) vision enables users (i.e. people, businesses, and systems) to allocate and consume the required computing services whenever and wherever they want in a context-aware seamless transparent manner. Hence, chosen services automatically interact with one another in a transparent manner to accomplish the required users’ goals. Middleware software plays an essential role in supporting such interactions, as it hides services heterogeneity and ensures their interoperability. Middleware enables services to locate one another without a priori knowledge of their existences and enables them to interact with one another even though they are running on different devices and platforms [1]. Services interactions are conducted via exchanging messages. A conversation message indicates the operation to be performed by the service receiving the message. A sequence of messages exchanged between services to achieve a common goal constitutes what is known by a conversation pattern. A set of conversation patterns is referred to as a service conversation. However, services may use different concepts, vocabularies,