On the Use of Similarity or Query Languages in Cloud Discovery Based on Ontology

On the Use of Similarity or Query Languages in Cloud Discovery Based on Ontology

Rawand Guerfel (ENIT, Université de Tunis El Manar, Tunis, Tunisia), Zohra Sbaï (ENIT, Tunis, Tunisia) and Rahma Ben Ayed (ENIT, Université de Tunis El Manar, Tunis, Tunisia)
DOI: 10.4018/IJSSMET.2017070104

Abstract

Cloud computing is increasingly used so that the number of providers offering services is rapidly increasing. Thus, a need to organize these services and to express relations between them arises. To answer this need, ontologies are used. To query these services, the authors use query languages, such as SPARQL, that return two types of results: either a list of required services, or an empty list. However, the second result is not desired. In fact, if the required service is not available, users want to be offered by a list of similar ones instead of the empty list. It is in this sense that the similarity, which provides more results ranked according to their utilities, is used. This paper first presents the Cloud ontology on which the authors' work is based. It then defines and compares between two Cloud service discovery methods which are: the discovery based on query languages and the discovery based on similarity. To show the efficiency of the search based on similarity, the authors propose a search engine that allows the users to query services using a simple to use interface.
Article Preview

1. Introduction

According to NIST (Mell & Grance, 2011) Cloud computing provides scalable IT capabilities that are offered as a service over the internet to multiple users. These users share common IT resources, reducing costs and resulting in greater computing efficiency. Cloud computing offers scalability that allows them to scale these capabilities and process power on-demand. It also offers a computer power to users when they need it by just turning these needs. Last but not least, the Cloud allows them to save their money. Indeed, users only pay for what they used and not for having equipment sit around (Padilla et al., 2015).

There are basically three layers to the Cloud that are used differently based on what they offer. The first layer is Infrastructure as a Service (IaaS) which offers virtual systems which can be connected using internet. In this layer, users can install software and even services. The second layer is Platform as a Service (PaaS) which is a proof model for running applications without the hassle of maintaining the hardware and software infrastructure at the company. The last layer is Software as a Service (SaaS) which is a delivering way of application as a service (Tiwari & Joshi, 2015). Using this layer, one is not obliged to install and maintain software (Guo & Zheng, 2015).

The Cloud computing is pretty big and is growing bigger every day. This caused the increase of the number of Cloud service providers as well as the number of services offered to end users. This increase in number makes the discovery mechanism a tedious task. It is in this context that many researchers have resorted to the use of ontology when researching for the appropriate service. In fact, ontology allows the sharing and the definition of a common understanding of the whole structure of information of different Cloud services offered by numerous providers. It organizes these services in a well-defined hierarchy and allows the developer to express relations between them. This facilitates the research task and makes it rapid and precise (Aloui & Touzi, 2015). Ontologies have been used for several years in knowledge Engineering (KE) and Artificial Intelligence(AI) to structure different concepts in a certain domain. The concepts are combined and considered as building blocks for expressing domain knowledge that they cover namely health domain, biology domain, etc. Cloud computing is also one of the most covered domain by the ontology (Spear, 2006).

To describe metadata in an ontology and facilitate their treatment, many languages, based on XML syntax, have been proposed such as RDF (Candan et al., 2001), RDFS (Nejdl et al., 2000), OWL-Lite (Euzenat & Valtchev, 2006), OWL-DL (Antoniou & Harmelen, 2004), etc. These languages give context to manipulate and structure the knowledge of the metadata. They are accessible through queries using query languages such as SPARQL (Kostylev et al., 2015). Indeed, by executing these queries, we can have two types of results: either a list of services having been requested by the user, or an empty list that does not contain services because of their non-existence. So, the drawback of query languages is that, in case of the non-existence of the required service, they do not search for similar ones that can meet the user demand. This drawback is resolved by the research based on similarity. In fact, the similarity makes from the Cloud service research a more performant task. It offers to the user more choices so he can select the appropriate service.

In summary, our approach consists of two main points. First of all, we present the ontology that we have developed using OWL. Then, we propose a search method for Cloud services discovery based on ontology that allows the user to write its query using an easy to use interface. Once written, this query was first converted to SPARQL language and executed on the designed ontology. In a second step, we improve this engine so that it is not anymore based on a query language. In fact, it uses the similarity mechanism when researching for the required service. Then, a comparative study between the first and the second solution is performed to show the importance of the similarity in the Cloud service discovery mechanism based on ontology.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 10: 4 Issues (2019): Forthcoming, Available for Pre-Order
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing