Section: Application Domains

Service Oriented Architectures (SOA)

Service Oriented Architectures aim at the integration of distributed services and more generally at the integration of distributed and heterogeneous data, at the level of the Enterprise or of the whole Internet (big data dimension).

The team seeks solutions to the problems encountered here, with the underlying motivation to demonstrate the usefulness of a large-scale distributed programming approach and runtime support as featured by ProActive and GCM:

• Interaction between services: the uniform usage of web services based client-server invocations, through the possible support of an Enterprise Service Bus, can provide a simple interoperability between them. For more loosely coupled interactions between services (e.g. compliant to the Web Services Notification standard), we pursue efforts to support publish-subscribe interaction models. Scalability in terms of number of notified events per time unit, and full interoperability through the use of semantic web notations applied to these events/data are some of the key challenges the community is addressing and we too. Events also correspond to data that may be worth to store, for future analytics, besides being propagated to interested parties (in the form of the event content). Our research can thus also contribute to the Big Data domain: we started to focus on how the use of flexible distributed and reconfigurable programming approaches through software components can allow us to devise powerful and flexible analytics on big data flows.

• Services compositions on a possibly large set of machines: if service compositions can even be turned as autonomic activities, these capabilities will really make SOA ready for the Open Internet scale (because at such a scale, a global management of all services is not possible). For service compositions represented as GCM-based component assemblies, we are indeed exploring the use of control components put in the components membranes, acting as sensors or actuators, that can drive the self-deployment and self-management of composite services, according to negotiated Service Level Agreements. For service orchestrations usually expressed as BPEL like processes, and expressing the composition in time aspect of the composition of services, supports for deployment, management, and execution capable to support dynamic adaptations are also needed. Here again we believe a middleware based upon distributed and autonomous components as GCM is really helpful.