Massive Data Streams Flow from Wall Street

Automated high-speed financial trading surged in volume in 2007, and now represents 22 percent of all trading, according to Larry Tabb, CEO of TABB Group. Meanwhile, the volume of financial trading handled by brokers placing orders over the phone fell from 47 percent in 2006 to 37 percent in 2007.

Tabb, whose firm tracks the impact of technology on the capital markets, says this trend of "disappearing traders" is reshaping the infrastructure that powers the US financial markets. "We're seeing a tremendous shift from high-touch channels to low-touch channels," Tabb said Tuesday at the O'Reilly Money:Tech conference in New York. "There's a tremendous amount of technology required to execute these trades. The order flow is becoming increasingly electronic."

That shift has generated demand for data center space from Wall Street firms and hedge funds, as well as high-speed networking infrastructure to execute low-latency trades. Savvis (SVVS) and Equinix (EQIX) have each seen strong demand from financial firms for colocation space in their New York and Chicago data centers.

In low latency trading, Wall Street firms use computerized algorithm-based trading systems to execute trades in milliseconds and even microseconds (one-millionth of a second). Ultra-low latency trading, also known as proximity trading, allows traders faster connections by placing their equipment as close as possible to the servers of the exchange executing the trade.

Timely market information is crucial in algorithmic trading. Last summer Savvis introduced a service to provide its financial customers with enhanced access to data feeds from major exchanges.

"There's been a tremendous growth in quote traffic," said Tabb. "If you're in this game, you need to gather real-time data from at least three major exchanges, and five regional exchanges," as well as electronic communication networks (ECNs) and private trading operations known as "dark pools."