Skeleton Automata for FPGAs: Reconfiguring without Reconstructing
Jens Teubner, Louis Woods, and Chongling Nie
Proceedings of the 2012 ACM SIGMOD Conference on Management of Data, Scottsdale, AZ, USA, May 2012.
While the performance opportunities of field-programmable gate arrays (FPGAs) for high-volume query processing are well-known, system makers still have to compromise between desired query expressiveness and high compilation effort. The cost of the latter is the primary limitation in building efficient FPGA/CPU hybrids.
In this work we report on an FPGA-based stream process- ing engine that does not have this limitation. We provide a hardware implementation of XML projection that can be reconfigured in less than a micro-second, yet supports a rich and expressive dialect of XPath. By performing XML projection in the network, we can fully leverage its filtering effect and improve XQuery performance by several factors.
These improvements are made possible by a new design approach for FPGA acceleration, called skeleton automata. Skeleton automata separate the structure of finite-state au- tomata from their semantics. Since individual queries only affect the latter, with our approach query workload changes can be accommodated fast and with high expressiveness.
camera-ready for SIGMOD 2012
submission to SIGMOD 2012 (conditional accept; with shepherding)
submission to VLDB 2011 (rejected)
submission to SIGMOD 2011 (rejected)