Buscar

Estamos realizando la búsqueda. Por favor, espere...

Mortadelo: automatic generation of NoSQL stores from platform-independent data models

Abstract: In the last decade, several NoSQL systems have emerged as a response to the scalability problems manifested by classical relational databases when used in Big Data contexts. These NoSQL systems appeared first as physical-level solutions, initially lacking any design methodologies. After this initial batch of systems, several design methodologies for NoSQL have been recently created. Nevertheless, most of these methodologies target just one NoSQL paradigm. In addition, as each methodology uses a different conceptual modeling approach, NoSQL database designers would need to remake conceptual models as they switch from one NoSQL paradigm to another. Moreover, most of these design processes provide just a set of design heuristics and guidelines that database designers need to apply manually, which can be a time-consuming and error-prone process. To overcome these limitations, this article presents Mortadelo, a model-driven NoSQL database design process where, from a high-level conceptual model, independent of any specific NoSQL paradigm, an implementation for a concrete NoSQL database system can be automatically generated. Moreover, this database generation process can be customized, so that some design trade-offs can be managed differently according to each context needs. We evaluated Mortadelo?s capabilities by generating database implementations for several typical NoSQL case studies. In these cases, Mortadelo was able to generate implementations for the Cassandra and MongoDB NoSQL systems from the same conceptual data model. These implementations were similar to the ones generated by design methodologies specifically developed for a single paradigm. Therefore, design quality is not sacrificed by our approach in favor of generality.

 Autoría: de la Vega A., García-Saiz D., Blanco C., Zorrilla M., Sánchez P.,

 Fuente: Future Generation Computer Systems, 2020, 105, 455-474

 Editorial: Elsevier

 Año de publicación: 2020

 Nº de páginas: 20

 Tipo de publicación: Artículo de Revista

 DOI: 10.1016/j.future.2019.11.032

 ISSN: 0167-739X,1872-7115

 Proyecto español: TIN2017-86520-C3-3 R

 Url de la publicación: https://doi.org/10.1016/j.future.2019.11.032