Financial market data processing solutions minimize UDP latency via modular software and C-to-FPGA acceleration

KIRKLAND, USA: Impulse and Luxoft today announced a strategic partnership to develop and deliver accelerated solutions for low-latency financial processing. The partnership leverages engineering talent from the US and Russia, and uses the latest advances in modular, highly adaptable software and reconfigurable hardware.

Impulse and Luxoft created powerful software modules for implementing building block functions in low-latency financial processing. These functions include real-time financial feed handling, searching for specific ticker symbols, comparing real-time data to historic data, and placing information-based orders in the minimal possible time.

The modular construction of this software enables users to insert their own analytic intellectual property, in-line, without the need to divulge their proprietary IP to 3rd parties. Impulse and Luxoft can provide anything from simple IP modules for specific processing tasks, to entire installed turn-key systems.

“Hardware independent software modules, and C-to-hardware compilation enables us to offer our clients the quickest, most cost-effective initial solution and, more importantly, an infrastructure that is easily adaptable to emerging hardware innovations,” explained Brian Durwood, Impulse CEO. “By using reconfigurable software modules, our clients can migrate to the newest hardware soon after it becomes available.”

“After watching design languages come and go over the years, we are seeing a return to basics. The Quantitative Analytical Engineers employed in most investment banks are turning back to C,” said Yuri Elkin, Managing Director, Financial Service, Luxoft. “Impulse C really covers the basics of adapting high-speed, microprocessor-oriented C to the more parallel world of FPGAs. This preserves the banks’ investment in C, and taps the world’s largest library of existing IP.”

The Luxoft/Impulse collaboration is an example of hardware/software co-design, in which computational resources are optimized for each specific software problem. This design method leverages highly reconfigurable, highly parallel computing devices called FPGAs.

FPGAs offer high performance and high data throughput, while operating at slower clock speeds and using less power than microprocessors. FPGA devices are increasingly important in applications from high-throughput video processing to cryptography and genomics. FPGAs are also excellent platforms for low-latency financial computing, with proven deployments in hedge funds and major banks.