Author:
Islam Elgedawy
Affiliation:
Middle East Technical University and Northern Cyrpus Campus, Turkey
Keyword(s):
Service Discovery, Self-configuring Services, DISCO, JAMEJAM, Service Knowledge Management.
Abstract:
The service discovery process involves many complex tasks such as service identification, composition, selection, and adaptation. Currently, there exist many discovery schemes that separately handle such discovery tasks. When a company needs to build a discovery service, it manually selects the suitable discovery schemes, encapsulates them as services, then invokes them as a composite web service. However, when different discovery tasks/schemes are needed, such composite discovery service needs to be manually reconfigured, and different versions of the discovery service are created and managed. To overcome such problems, we propose to build a dynamic self-configuring discovery service (i.e., DISCO), that takes the required discovery policy from users, then automatically finds the suitable discovery schemes in a context-sensitive manner, and finally arranges them as a collection of executable BPEL processes. This is done by adopting different types of knowledge regarding the services’
aspects, discovery schemes, and the adopted software ontologies. Such different knowledge types are captured and managed by the previously proposed JAMEJAM framework. Experimental results show that DISCO successfully managed to reconfigure itself for different discovery policies.
(More)