¿Es posible la convergencia HPC y Big Data? Lecciones aprendidas de los sistemas de E/S

Prof. María S. Pérez gave an invited talk on April 11 at UCM.

The HPC and Big Data (BD) areas have followed different paths due to the existence of two divergent communities and the different objectives pursued by both. However, in recent years there has been a change in both HPC and BD applications. On the one hand, the HPC applications use an increasingly high volume of data, requiring many times the ability to visualize and analyze such data. On the other hand, BD applications need greater computational capacity, due to more ambitious challenges and the combination of the analysis of data with simulation processes. The existence of the so-called Extreme Data Analytics creates the need to combine solutions oriented both to improve access and data processing as well as computing. At the storage level, we have carried out work that deals with such convergence. This may constitute a first step that could be extended to other characteristics of computer systems, and therefore, achieve the desired convergence between HPC and BD.

The slides can be found here.