FPGA-based hardware acceleration for a key-value store database

Abstract

As the modern world processes larger amounts of data and demand increases for better transaction performance, a growing number of innovative solutions in the field of database technologies have been discovered and implemented. In software and systems design, the development and deployment of the NoSQL (Not Only SQL) database challenged yesterday's relational database and answered the demand for exceedingly higher volumes, accesses, and types of data. However, one less investigated route to bolster current database performance is the use of dedicated hardware to effectively complement or replace software to 'accelerate' the overall system. This thesis investigates the use of a Field-Programmable Gate Array (FPGA) as a hardware accelerator for a key-value database. Utilized as a platform of reconfigurable logic, the FPGA offers massively parallel usability at a much faster pace than a traditional software-enabled database system. This project implements a key-value store database hardware accelerator in order to investigate the potential improvements in performance. Furthermore, as new technologies in materials science and computer architecture arise, a revision in database design welcomes the use of hardware for maximizing key-value database performance.