The Glass Half Full: Using Programmable Hardware Accelerators in Analytics

Even though there have been a large number of proposals to accelerate databases using specialized hardware, often the opinion of the community is pessimistic: the performance and energy efficiency benefits of specialization are seen to be outweighed by the limitations of the proposed solutions and the additional complexity of including specialized hardware, such as field programmable gate arrays (FPGAs), in servers. Recently, however, as an effect of stagnating CPU performance, server architectures started to incorporate various programmable hardware components, ranging from smart network interface cards, through SSDs with offloading capabilities, to near-CPU accelerators. This availability of heterogeneous hardware brings opportunities to databases and we make the case that there is cause for optimism. In the light of a shifting hardware landscape and emerging analytics workloads, it is time to revisit our stance on hardware acceleration. In this paper we highlight several challenges that have traditionally hindered the deployment of hardware acceleration in databases and explain how they have been alleviated or removed altogether by recent research results and the changing hardware landscape. We also highlight that, now that these challenges have been addressed, a new set of questions emerge around deep integration of heterogeneous programmable hardware in tomorrow’s databases, for which answers can likely be found only in collaboration with researchers from other fields