FXSpotStream to Deploy Microsecond Architecture
Posted by Colin Lambert. Last updated: October 6, 2021
FXSpotStream has announced the deployment of a new low-latency architecture, which will be deployed globally, and, the firm says, deliver substantial improvements to the service’s market data processing times. The project is already underway in the New York site and will involve a full overhaul of the existing infrastructure in place for FSS’ liquidity providers and clients.
Once in production, starting Q4 with a phased roll out in New York followed by London and Tokyo in Q1 of next year, market data processing times under all market conditions are targeted to not exceed a maximum of 250 microseconds, the firm says. Additionally, the new architecture will provide scalability to allow the service to maintain the same performance levels as FSS continues to grow.
“We have grown tremendously since our launch in 2011 becoming the 3rd largest FX Service by volume with an ADV this year at $48.5 billion,” says FXSpotStream co-founder and CEO, Alan Schwarz. “At the same time, we have placed great importance on the speed and quality of our market data distribution. Our aim is to provide our clients plus liquidity providers the best-in-class infrastructure and prioritize the performance and reliability of the service in the same way as our deep product offering. Ensuring that our clients receive pricing and can send orders in a timely manner is of critical importance.
“We have set a baseline targeting that clients’ market data latency in all market conditions, including during peak periods, will not exceed 250 microseconds,” he continues. “As a result, we will see times that are much faster than this target, but we are setting a maximum time to provide a level of consistency in performance during periods of high market volatility.
“As we look forward to the next decade, we remain very bullish about our business,” he concludes. “We recently launched over our API support for all of our banks’ algos and supported allocations functionality. In just two months since we went live with our algo offering we have already supported $3.2 billion of client algo orders. Work to add a GUI to support algos and allocations is underway and is expected to be live by late Q1 2022.”