Avalanche - Stream Processing on Bare Metal
Avalanche was founded in 2008 by Jens Teubner as a research project on the use of field-programmable gate arrays (FPGAs) in a database context. The project was funded by one of the prestigeous Ambizione grants of the Swiss National Science Foundation from 2009 to 2013, when Jens Teubner was at ETH Zürich.
The project pioneered the use of FPGAs for database acceleration. Glacier is a query compiler that translates a rich subset of SQL into equivalent hardware circuits. This allows stream processing at guaranteed network speed and thus removes a bottleneck that had hampered many important application classes (such as high-frequency trading, HFT) before.
FPGAs demand to formulate application problems in massively parallel ways. We devised ways to do so, e.g., for complex event processing (CEP), frequent item counting, and skyline computation. A guiding principle, thereby is an awareness of the communication patterns among parallel units in the hardware circuit.
Avalanche students:
- Pratanu Roy (PhD student at ETH)
Avalanche graduates:
- Louis Woods (PhD, graduated 2014 from ETH, now with Oracle Labs)
- René Müller (PhD, graduated 2010 from ETH, now with IBM Research)