The alliance allows Orc clients, including market makers and volatility traders, to run algorithmic trading strategies based upon field programmable arrays (FPGA) data feeds from NovaSparks, which operate at nanosecond processing speeds.
The FPGA-based market data feeds are said to be “deterministic” because they maintain a consistent speed regardless of data rates or number of updates, number of markets watched or number of users, said a spokesperson for NovaSparks. As the number of updates gets higher, NovaSpark’s Deterministic FPGA Matrix architecture has the ability to process them in parallel so there is no queuing of messages.
Orc said it conducted extensive research to determine the benefits of FPGA technology to market makers and volatility traders who are constant facing technology risk around speed. Customers that take advantage of this alliance will experience more predictable speed for market data and overall improved performance trading performance, said Orc.
“In our beta test phase, clients told us they saw an increase in trading profitability,” commented Joakim Dahlstedt, CTO at Orc in a statement. “For some of our clients, determinism of high speed market data is even more important than pure speed,” added Dahlstedt, CTO Orc.
Additionally, since market data is processed on the FPGA appliance and not on the Orc servers, the server CPU is able to offer more processing power to Orc Liquidator, resulting in additional latency benefits, the company said.
“Market data feeds that match the speed of the market, even during market spikes, is crucial for the profitable operations of traders,” stated Yves Charles, CEO of NovaSparks. The company's feed handlers are currently deployed in major North American and European derivatives and equities markets, according to its web site.
The move is important for traders who make most of their money in a high-volume market, noted the spokesperson. If market data volumes spike, and the timeliness goes down, that is when algo traders can lose money. With consistent market data, their strategies should work in low or high volume markets.
Dahlstedt said that Orc’s customers can adopt the deterministic market data feed “with 100 percent compatibility to their current trading algos.” The company said that clients could adopt this high-speed deterministic market data infrastructure without any modifications to their current algorithms or Orc products.
Both companies are exhibiting their products at FIA Expo in Chicago this week from Oct. 30- November 1.