Over the last year, Swarm64 has been rolling out FPGA-accelerated databases to customers. Those customers are happy and with that encouragement, we are aiming for big roll-outs this year.
With success comes responsibility, including transparency and open communication to the world about us, our product, our mission and our values.
When we conceived the first version of the Swarm64 Scalable Data Accelerator (Swarm64 DA), we were quite naïve about the size of the challenge we undertook. Today, we better understand why experts consider acceleration of existing databases transparently to the application layer the toughest challenge and the most valuable contribution to big data analytics.
With PostgreSQL, Swarm64 DA interfaces through the Foreign Data Wrapper interface. The SDA is designed for big and fast data analytics. It complements the transaction processing capabilities of PostgreSQL. Tables can be shared freely between the transactional engine and the SDA. We accelerate data analytics by arranging data differently and such that its processing can be shared between the FPGA and the CPU very efficiently. It is a complete solution stack and the FPGA is hidden in the middle of it. Customers can not access the FPGA and actually, do not even need to know that it exists.
So far, we covered only a small part of the solution space and already with these achievements, we deliver significant value to our customers. If you share our excitement for database technology, hardware design, and creating value for database users, please come back to this blog!