Ever-increasing volumes of data in the US market are posing a challenge to the buy-side, as it attempts to speed up its routing decisions. But Barry Thompson, founder and chief technology officer of US messaging systems provider Tervela, believes there are solutions to this mounting problem.
In the US trading environment, there is a vast array of execution venues. Coupled with this, trading volumes are skyrocketing, and there is a widespread and growing use of electronic trading techniques such as execution algorithms and automated alpha models. There are also finer tick sizes to contend with, following the decimalisation of the US equities market in 2001 and the options market in 2007.
The result is a flood of market data that buy-side traders need to store and access to trade effectively. Traders need to process this increasing volume of data quickly so they can make split-second trading decisions. But Barry Thompson, founder and chief technology officer of US messaging systems provider Tervela, believes new technologies already in the pipeline could help these traders get up to speed.
“The volume of market data that the buy-side has to consume now is immense,” says Thompson. “A buy-side firm, depending on its size, may have to take an entire data feed from the Options Price Reporting Authority, even though it only needs data for 20 or 30 options symbols across a few venues. A hedge fund might have an automated trading model that needs very large feeds from venues such as NYSE Arca, but it is not necessarily cost-effective for them to take entire feeds.”
He claims there are a number of new offerings that could grant traders fast access to the data they need most. While technology vendors are creating high-performance data feed handlers and messaging solutions, exchanges are offering filters to help make the flow of data more manageable for traders. “When you connect to the exchanges, rather than taking a raw data feed, you can get an application programming interface that provides filtered market data,” explains Thompson. “If you only need 50 symbols, you only get the market data on those 50 symbols.”
When employing new technologies such as feed handlers, Thompson says buy-side firms need to consider factors such as total cost of ownership – including the number of machines required to manage the feed handlers and the cost of upgrades to them. As well as inevitable integration considerations, buy-side firms should also look at how well different feed handlers interoperate. “For example, if I buy one feed handler today and I need another one tomorrow, what will my cost be to change? And, what if I mix and match – getting seven feeds from one vendor and three from another one?” says Thompson.
Getting access to the most relevant data is only part of the battle. Traders need to process the data and make immediate order routing decisions based on it. Thompson contends that there is little point having sub-millisecond data feeds if trading systems are taking between 10 and 15 milliseconds to decide where to send orders.
As trading latency becomes more critical, initiatives being developed to help speed up routing. “Emerging technologies will provide low-latency FIX engines and high-performance solutions on the order routing and intelligent liquidity discovery side so you know which agent to clear through or which liquidity venue to route those orders to,” he says.
However, not all traders are aware that there are products available to help them bear the increasing data load. “I think the buy-side needs to learn a lot more about this,” says. Thompson “There is a burgeoning awareness among the highly-automated shops that they don’t have to build everything themselves, but not among the more traditional buy-side firms.”