DaaS Repository through MIDAS Web Crawler

Abstract

Over the years, SaaS (Software-as-a-Service) and DaaS (Data-as-a-Service) have respectively become common models for application delivery and cloud data delivery. However, the heterogeneity of these services makes it difficult to automate this process. In these cases, a solution for interoperating SaaS and DaaS is required. MIDAS (Middleware for DaaS and SaaS) is a proposal that does not present any mechanism that dynamically searches and updates DaaS, although it allows interoperating these layers. This limitation of MIDAS compromises interoperability, as it may result in outdated or non-existent DaaS. Therefore, it is necessary to have a repository inside the MIDAS, called DIS (Dataset Information Storage), to store DaaS metadata in the clouds. However, DaaS information must be registered in an automated and periodic manner, in order to guarantee the traceability and updating of DaaS. Thus, the present paper proposes the development of a Web Crawler to be incorporated into the MIDAS middleware, to (i) periodically update the DIS information, guaranteeing the consistency of DaaS, and (ii) analyze the existence of semantic similarities among DaaS, so that the DaaS are aligned semantically. Experiments were carried out with the purpose of validating the proposed work, in order to evaluate correctness, recall, execution time, overload, ontological consistency, and correction of alignments. The results obtained demonstrate the effectiveness and efficiency of our proposal.