FIX Group to Promote FAST Protocol for Latency Communications
January 11, 2010
A new committee of the FIX Protocol Group in London will likely advocate that latency monitoring vendors use the FIX Adapted for Streaming protocol as a language for communicating information on latency rates with each other.
Introduced by FPL’s Market Data Optimization Working Group in 2005, the FAST protocol allows for the high-speed delivery of market data with low bandwidth. Users include NYSE Arca, the BATS Exchange and the Chicago Mercantile Exchange.
The new committee is the FIX Inter-Party Latency Working Group, created in November and planning to hold its first meeting by conference call at 8 a.m. Eastern time on Wednesday, January 13.
“The goal of the group is to develop standards for measuring and reporting on latency across multiple organizations so that firms can develop a better understanding of where their latency hot spots exist and how they can address them,” says Henry Young, co-chair and director of product development at TS-Associates, a latency monitoring provider in London. The other co-chair is Nicholas Morrison, head of market data and middleware technology at Nomura Europe in London.
Inter-party latency refers to the time it takes to communicate information on market data and execute trade orders between a firm and its trading venue rather than through its internal applications. Higher latency rates can cost multi-millions in lost revenues to a brokerage firm involved in high-frequency and arbitrage trading; such firms talk in sub-millisecond rather than sub-second targets.
“To understand inter-party latency, the first requirement is to monitor latency at all the significant points of a financial transaction, and the second requirement is to join the dots and understand which messages – in a variety of message formats with different IDs and tags – actually comprise a single financial transaction,” says John Barr, research director for financial markets and head of European Union Research for The 451 Group, a New York based technology research firm.
While latency monitoring vendors are all trying to measure end-to-end latency and expand the types of financial messaging protocols they can interpret, they do not use a consistent set of metrics or communication protocols. That means that if a firm wants to understand the latency rate between his or her firm and its trade execution and market data venues—as well as the cause of the latency -- it must license multiple latency monitoring tools. That means the same ones used by the trading venues themselves.
As trading volumes and volatility have grown over the past year, latency monitoring service providers have expanded into the inter-party latency arena. Among the top providers measuring network or financial message data cited by Barr are Correlix in New York, Corvil in Dublin and TS-Associates in London.
“Since interparty latency is a relatively new area, it is too early to say who the winners and losers will be, but we expect to see lots of action in 2010,” says Barr “TS-Associates took the farsighted view that better interparty latency monitoring – even if it means collaborating with competitors – is good for the industry because it helps to grow the number of firms engaging in latency analysis.”
Young doesn’t believe that standardization between latency monitoring service providers won’t stymie competition because they can still compete on the depth of their analysis, the range of messaging protocols they support and their ability to extend their products beyond measuring interparty latency to exchange infrastructures and trading infrastructures.
