qSpark Developer Talks Low-Latency, High-Frequency Trading Infrastructure

In an interview with Benzinga, Eliezer Paiewonsky, director of DevOps at qSpark, a provider of low latency trading platforms, discussed progress on a field-programmable gate array solution to help bolster risk management and high-speed executions.

About qSpark

qSpark is a tech company that provides advanced trading platforms that enable low latency execution of proprietary algorithms and strategies via co-located infrastructures.

“We’ve basically developed an in house, end-to-end solution for trading; so, everything from the market data handlers, to the interfaces of the exchanges, to actually sending orders, and the full risk management system that kind of sits in front of your connections,” Paiewonsky said. 

That's in addition to execution-side services, or infrastructure that enables communication with both market data and order entry. qSpark provides the broker, or clearing firm, with what's needed to monitor trading, including end-of-day files, consolidated audit trails, archiving and so on. 

“It truly is a one-stop shop for both … the broker and trader,” the exec said. 

Technologies And Connectivity

qSpark is a registered service bureau that is connected with all of the major exchanges, dark pools and counter parties.

“From an execution perspective, we are connected to pretty much everyone or at least anyone major,” Paiewonsky said. 

qSpark has a full C++ API; regardless of underlying protocol language, “you have one standard API you can use to easily talk to any of the exchanges,” he said.

In regards to new tech, qSpark is spearheading development on a field-programmable gate array, or FPGA: an end-to-end solution for low-latency trading that routes incoming market data through an algorithm that is "burned, synthesized, and ultimately loaded onto the chip, along with a full risk layer," Paiewonsky said. 

Adaptability

According to Paiewonsky, qSpark is constantly evaluating its approach to latency and throughput.

“This industry is ever-changing; different venues are changing their protocols — for instance, NYSE recently migrated from … legacy to pillar protocol," he said. "That is the kind of good thing going through someone like us — we handle all of the connectivity and coding for you." 

The company is working to improve layer-switching infrastructures. The standard for network communication is approximately 300 to 350 nanoseconds; qSpark is working to drop “that down to single and double digits,” Paiewonsky said. 

Moving forward, the firm aims to work more with technologies like machine learning to develop a solution that allows traders to derive actionable insights from “real time and historical” data sets, he said. 

Related Links: 

Pico Acquires Corvil In Deal That Aims To Provide 'Tremendous Value To Clients'

SmartBotCoin Unveils 'World's Most Advanced' Cryptocurrency Trading Platform