A DHT-based semantic overlay network for service discovery

• Published in: Future generation computer systems Vol. 28; no. 4; pp. 689 - 707
• Main Authors: Pirrò, Giuseppe, Talia, Domenico, Trunfio, Paolo
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.04.2012
• Subjects: Distributed hash tables; Semantic overlay networks; Semantic similarity; Semantic Web services; Distributed hash tables; Semantic similarity; Semantic Web services; Semantic overlay networks
• ISSN: 0167-739X; 1872-7115
• DOI: 10.1016/j.future.2011.11.007

The number of available Internet services increases every day. This trend demands distributed models and architectures to support scalability as well as semantics to enable efficient publication and retrieval of services. Two common approaches toward this goal are semantic overlay networks (SONs) and distributed hash tables (DHTs) with semantic extensions. SONs enable semantic-driven query answering but are less scalable than DHTs, which, in their turn, feature efficient but semantic-free query answering based on exact match. This paper presents a strategy and a system, called ERGOT, that combine DHTs and SONs to enable semantic-based service discovery in distributed infrastructures such as Grids and Clouds. ERGOT uses semantic annotations that enrich service specifications in two ways: (i) services are advertised in the DHT on the basis of their annotations, thus allowing us to establish a SON between service providers; (ii) annotations enable semantic-based service matchmaking, using a similarity measure between service requests and descriptions. An extensive evaluation of the system is presented and discussed. The experimental evaluation we carried out confirmed the efficiency of the implemented strategy in terms of both accuracy of search and network traffic. ► ERGOT combines DHTs and SONs for semantic service discovery in distributed systems. ► Annotated services are advertised on a DHT, building a SON between service providers. ► Similarity between requests and descriptions enables semantic service matchmaking. ► Experimental results show that ERGOT leads to high search accuracy with low overhead.