CASPIAN JOURNAL

MANAGEMENT AND HIGH TECHNOLOGIES

MINING FOR STABLE SITUATIONAL SERVICE COMPOSITIONS

Read

Ishkina Evgeniya G.  MINING FOR STABLE SITUATIONAL SERVICE COMPOSITIONS // Caspian journal : management and high technologies. — 2011. — №3. — pp. 12-19.

Ishkina Evgeniya G. - Post-graduate student, Astrakhan State University, 20a Tatishchev str., Astrakhan, 414056, Russia, ishkina@aspu.ru.

This paper describes the collective services memory component inside the self-adaptive middleware for mobiquitous systems in the framework of ASTRA project (Adaptive Service Technologies for Reliable Access). The main idea of our middleware approach is to use collective service intelligence for providing users with customized services which are the most appropriate for concrete usage situations in different domains (tourism, education, etc.). The core of ASTRA approach consists of a middleware supporting three key peculiarities: collective domain-independent service management; task-awareness as main factor of interaction; and finally, ability of self-adaptation, i.e. possibility to find new service compositions and criteria of their appropriate usage. In this paper we describe the structure of collective services memory component and data flows between it and other middleware components, namely service integration component (atomic services description flow), usage analysis component (situational service annotations flow), service mining component (useful service compositions flow), and situational service composition component (flow of appropriate services for the given situation and flow of services implicit relationships). Each data flow is described in details along with related algorithms.

Key words: service computing,service-oriented architecture,adaptive services,service composition,service mining,data mining,situation awareness,context awareness,ASTRA project,middleware,task-aware computing,goal-driven service composition.