Domain Objects for Dynamic and Incremental Service Composition

Abstract

Abstract. A key feature of service-based applications (SBAs) is the capacity to dynamically define the composition of services independently available, which is required to achieve user goals. For this reason, to effectively deal with the obstacles due to continuous context changes, an incremental refinement of provided services is needed. We propose a new model that allows service functionalities to be defined partially, through the use of abstract activities. The refinement of these activities is postponed and performed incrementally at runtime, using the actual context as a guide. Our approach lets a service provider avoid the hard-coding of all service functionalities and their possible compositions at design time, delaying their refinement until the execution phase. Consequently the whole SBA's design becomes modular and flexible to better meet the typical dynamism of SBA. We illustrate the approach through an example scenario from the urban mobility domain.

Design-Time to Run-Time Verification of Microservices Based Applications

“…The approach presented in [21] aims at dealing with environmental dynamism in service-based applications. The proposed modeling approach allows for partial definition of services in order to perform late (i.e., at runtime) incremental composition, when the execution context is discovered.…”

GraphEvol: A Graph Evolution Technique for Web Service Composition

“…a directed acyclic graph (DAG), stipulating the sequence in which each atomic service should be executed, as well as the output-input connections between services. Many approaches to Web service composition have been proposed in the literature, from variations on AI planning techniques [7,4] to the employment of integer linear programming solvers [1,23]. In particular, promising results have been achieved with the use of Evolutionary Computation (EC) techniques [21,19], because their search methods successfully handle the large search space characteristic of the composition problem.…”

Evolutionary Computation for Multifaceted Web Service Composition

“…Approaches for semantically binding concrete services to the abstract workflow include the use of engines such as METEOR-S [2] and path finding algorithms within a graph structure [167]. Other works focus on semantic descriptions of different aspects of the problem, including the specification of user requirements [104], the identification of suitable services within the repository [171], the ontological classification of QoS aspects of services [189], and the modelling of information needed in the composition workflow [96,122,27].…”