Abstract
FPGAs are promising target architectures for hardware acceleration of database query processing, as they combine the performance of hardware with the programmability of software. In particular, they are partially reconfigurable at runtime, which allows for the runtime adaptation to a variety of queries. However, reconfiguration costs some time, and a region of the FPGA is not available for computations during its reconfiguration. Techniques to avoid or at least hide the reconfiguration latencies can improve the overall performance. This paper presents optimizations based on query look-ahead, which follows the idea of exploiting knowledge about subsequent queries for scheduling the reconfigurations such that their overhead is minimized. We evaluate our optimizations with a calibrated model for various parameter settings. Improvements in execution time can be “calculated” even if only being able to look one query ahead.
Highlights
Some research has been conducted to address the acceleration of database query processing with the help of FPGAs, e.g. [16, 21, 27], and its integration into a DBMS
The paper is structured as follows: In Section 2, we introduce the concepts of the ReProVide approach and its ReProVide Processing Unit (RPU)
This model is used for evaluating the presented optimization strategies with values determined by measurements on the existing prototype of the RPU
Summary
The ReProVide (for “Reconfigurable data ProVider”) approach [5] is part of it. It introduces a storage-attached hardware technology called ReProVide Processing Unit (RPU) to process queries completely or at least partially close to the data source, in synergy with a DBMS. This co-design approach exploits dynamic reconfiguration of FPGAs in the RPUs in combination with a novel DBMS optimizer. On the RPUs, a library of query-processing modules is stored, which can be configured onto the FPGA in a few milliseconds. At each point in time, only a subset is ready for use in query
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
