Enhancing Semantic Interoperability on MIDAS with Similar DaaS Parameters.

The vast amount of social data and the extensive use of smart devices have enabled different formats of DaaS (Data as a Service). Cloud consumers are encouraged to access such DaaS from a SaaS (Software as a Service) directly. However, DaaS can evolve and modify some of its parameters, thus avoiding a SaaS to catch all such adjustments. Thus, it is important to have a middleware to bind the request from SaaS applications to data from DaaS. One such middleware is MIDAS (Middleware for DaaS and SaaS), which provides a way from SaaS to DaaS seamlessly. Although there are many advantages to publishing data as a DaaS format, the major challenge is the dynamicity of DaaS parameters, which can change and evolve. Such changes can affect SaaS applications, even providing runtime errors. Thus, our approach overcomes this problem by ensuring a transparent mechanism to DaaS parameters when modifications have occurred. Our work aims to recognize each similar attribute seamlessly, increasing availability to interoperate SaaS and DaaS through MIDAS. We evaluate the distance measures thought 22 parameters from 11 DaaS providers into five scenarios on parameters change. Two of them were the most outstanding measures (Cosine and Jaccard), and we included them in our approach to better obtain the similarity of two parameters. Each parameter was carried out toward WordNet thesaurus, and a second evaluation was performed analyzing our approach into MIDAS through three criteria: overhead, execution time, and correctness. Our experiments have shown that we are in a good direction to provide interoperability into SaaS and DaaS.