A genetic programming approach to support the design of service compositions

TLDR

Service‑oriented architectures are rapidly adopted, enabling software‑as‑a‑service and composite services that rely on service integrators to orchestrate multiple services. The authors aim to distinguish among user‑facing services to better understand and address challenges in service composition.

Abstract

Service-oriented architectures are rapidly gaining presence and popularity in the development of distributed software systems. Their rapid diffusion is changing the software engineering landscape and is opening the way to implement the idea of “software as a service” [19], i.e. something that one uses to compose a workflow that achieves a business goal, and then disengages. A service-oriented system can be thought of a set of services that collaborate in some way to carry on a given task. Also, service-oriented system can be, on its own, a service, usually referred as a composite service or composed service. A conceptual model for service-oriented systems [18] defines it as: “A Composed Service is a particular kind of Service, developed by a Service Integrator, which makes use of other Services”. From this definition emerges how the Service Integrator has a key role while dealing with service compositions. In fact, to have the right prospective of the existing problems, it is desirable to make a clear distinction between different services users i.e.: