Wall Street & Technology is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Trading Technology

10:46 AM
Connect Directly

NovaSparks Releases FPGA Feed Handlers for Five Equity Exchanges

The company's FPGA architecture achieves average latencies of 400 to 700 nanoseconds, avoiding bottlenecks by bypassing the CPU.

NovaSparks today announced the release of five market data feed handlers utilizing its FPGA Market Data Matrix architecture to process data from the key U.S. stock exchanges.

NovaSparks, based in Boston and Paris France, developed a matrix of FPGAs — which stand for field programmable gate arrays—embedded into the microprocessor chip.

As a result of speeding up the market data transmission rates the FPGA Feed Handlers process streaming data with an average latency of 400 nanoseconds and maximum latency of 700 nanoseconds, 99.99 percent of the time, the company stated.

The FPGA Feed handlers are now available for BATS PITCH for BYX and BZX feeds; Direct Edge NG-A and NG-X feeds; Arca XDP Integrated feed (available when the new format is released); and London Stock Exchange LSE Millennium Level 2 and Turquoise Millennium Level 2.

“Our FPGA Market Data Matrix Feed Handlers process the cash equity feeds in less than 700 nanoseconds even during micro bursts and message volumes peaks” stated Yves Charles, CEO NovaSparks, in today’s release. “No other commercial offering makes this claim”, asserted the CEO.

The feed handlers are designed for principal trading firms requiring consistent processing speeds, especially during heavy market volume. Referred to as deterministic latency, the FPGA Matrix architecture eliminates latency fluctuations typically seen in conventional software and hybrid FPGA approaches.

To speed up market data transmission rates, Novasparks uses a pure FPGA Matrix approach with no CPU involved, according to a spokesperson for the company. This is different from other market data providers that offload one or two functions to the FPGA chip, according to the spokesperson. Alternative approaches aimed at Wall Street using FPGAs, known as ‘hardware acceleration’, use FPGA cards inside traditional CPU based servers to offload a subset of functions from the CPU. “The NovaSparks approach is to use a matrix of FPGAs at the heart of the Feed Handler to eliminate all bottlenecks,” explained CEO Charles.

While other CPU-based systems clog up with high data rates and microbursts, which is a problem, said the spokesperson, NovaSparks doesn’t have those bottle necks which is why they have a consistent speed.

The company said it's rapidly rolling out feed handlers for equity and futures exchanges across Europe and North America. Feed handlers for CME, NASDAQ, and Eurex are currently available. NovaSparks is accelerating the development of the FPGA Market Data Matrix with proprietary compilers that lessen the need for VHDL programmers while tuning the firmware for consistent, ultra-low, nanosecond latencies.

The FPGA Market Data Matrix can scale across the entire market by chaining feed handler appliances together. Functionality of the feed handlers include subscription-based filtering, feed normalization and the ability to fan out the data using 10G or PCI (PCI Express) connections.

Ivy is Editor-at-Large for Advanced Trading and Wall Street & Technology. Ivy is responsible for writing in-depth feature articles, daily blogs and news articles with a focus on automated trading in the capital markets. As an industry expert, Ivy has reported on a myriad ... View Full Bio

Register for Wall Street & Technology Newsletters
Exclusive: Inside the GETCO Execution Services Trading Floor
Exclusive: Inside the GETCO Execution Services Trading Floor
Advanced Trading takes you on an exclusive tour of the New York trading floor of GETCO Execution Services, the solutions arm of GETCO.