Towards Combined Functional and Non-functional Semantic Service Discovery
Service-orientation is increasingly adopted by application and service developers, leading to a plethora of services becoming increasingly available. To enable constructing applications from such services, respective service description and discovery must
- PDF / 723,535 Bytes
- 16 Pages / 439.37 x 666.142 pts Page_size
- 37 Downloads / 192 Views
Abstract. Service-orientation is increasingly adopted by application and service developers, leading to a plethora of services becoming increasingly available. To enable constructing applications from such services, respective service description and discovery must be supported by considering both functional and non-functional aspects as they play a significant role in the service management lifecycle. However, research in service discovery has mainly focused on one aspect and not both of them. As such, this paper investigates the issues involved in considering both functional and non-functional aspects in service discovery. In particular, it proposes different ways via which aspect-specific algorithms can be combined to generate a complete service discovery system. It also proposes a specific unified service discovery architecture. Finally, it evaluates the proposed algorithms’ performance to give valuable insights to the reader. Keywords: Service · Discovery · Matchmaking · Semantics · Ontology · Performance · Evaluation · Functional · Non-functional · QoS · Architecture
1
Introduction
Nowadays, modern applications and business processes adopt service-orientation due to the many advantages it delivers, including loose coupling, re-usability, increased performance and cost reduction. To construct such applications, the services from which they are built need to be described appropriately, discovered and finally composed. Concerning service discovery, the state-of-the-art can be split into approaches that either focus on functional or non-functional aspects. Functional service discovery work [8] matches user’s functional requirements by exploiting various types of techniques from information retrieval and the semantic web [9,16]. Functional requirements and capabilities are mainly described via service IO while some work [7] covers behavioral aspects via service preconditions and effects which are however not available in real service advertisements. Most work nowadays exploits semantic techniques to exhibit better accuracy levels and other techniques to achieve performance speedup. However, the discovery accuracy is imperfect due to non-consideration of behavioural aspects. c IFIP International Federation for Information Processing 2016 Published by Springer International Publishing Switzerland 2016. All Rights Reserved M. Aiello et al. (Eds.): ESOCC 2016, LNCS 9846, pp. 102–117, 2016. DOI: 10.1007/978-3-319-44482-6 7
Towards Complete Semantic Service Discovery
103
Non-functional service discovery work [10] can be split into three main categories. Ontology-based approaches [19] employ subsumption techniques to infer the matching between ontology-based non-functional service descriptions but are suitable for unary-constrained specifications. Constraint-based approaches [5] exploit n-ary specifications as models, including quality terms drawn from common vocabularies, and particular metrics which involve solving one or more constraint (combined) models to infer the matchmaking. Mixed approaches [10] combine the best
Data Loading...
