CloudFNF: An ontology structure for functional and non-functional features of cloud services

Abstract

Recently, cloud computing becomes one of the main orientations of many researchers and companies in the IT area. Therefore, a huge number of cloud services have been developed. Because of the diversity and heterogeneity in providing these services, it is urgently needed to develop a unified cloud ontology. Such ontology can classify these services appropriately and participate as a mapping layer to present such services in a unified description format. Although many studies have been conducted to build cloud ontologies, they have adopted the cloud service layers-based structure. On the other hand, the existing cloud services may involve functionalities from different layers due to the continual increase in the complexity of customer demands. Unfortunately, there are no clear relations to organize the interventions among these layers. Therefore, such services may be difficult to be classified into a specific cloud service layer. Additionally, the layers-based structure of ontologies represents an obstacle to address important issues, such as cloud service recommendation. Despite there are few cloud service functionality-based cloud ontologies, they suffer from many overlaps, lack of semantic relations, or poor granularity of concepts. Also, all the existing cloud ontologies (i.e., layers-based and functionality-based) lack important criteria, such as completeness, consistency, conciseness, clarity, preciseness, and granularity. In this paper, a comprehensive cloud ontology called CloudFNF has been proposed to overcome such drawbacks. According to the structure of the proposed ontology, cloud services are classified as functionality-based instead of layers-based. Also, non-functional features of cloud services (i.e., configuration and QoS features) are considered to enable services of the same functionalities to be ranked efficiently. Based on our previously suggested cloud ontology evaluation framework, our proposed cloud ontology has been evaluated compared to the most related cloud ontologies. The evaluation results show that the proposed CloudFNF ontology outperforms the other ontologies.