One area that’s under cost-cutting scrutiny is market data infrastructure. Less than a decade ago, firms subscribed to data feeds from consolidators and used digital data distribution systems from the likes of Reuters, Tibco and Wombat to provide data across the trading floor.
As the popularity of algorithmic trading grew and required data with just milliseconds of latency, investment banks brought in raw data feeds and ticker plants to host that infrastructure, says Frank Piasecki, president of Activ Financial, a provider of market data systems. Brokerage firms set up unique ecosystems driving their own internal proprietary trading systems, he says, and replicated that market data infrastructure across the world and asset classes.
Today, the trading environment has changed. “IPOs are down, trading volumes are down and revenues are depressed, so firms are looking at the IT and maintenance costs as a drain on their budgets,” says Todd Albright, formerly Activ Financial’s senior VP and head of sales and marketing for the Americas, who has since left the company. They’ve built a lot of this themselves, while also growing through acquisitions, creating layered, competing infrastructures, he says.
Some investment banks are outsourcing market data systems “because it has gotten so complex and cumbersome to maintain,” Piasecki says. His firm is in discussions with multiple global investment banks that are looking to rationalize their market data infrastructure.
The market data burden goes beyond front-end development and maintenance; banks must also have enough capacity to handle the growth in market data message rates. “The automation of the markets means that quote volumes have gone up, driven by regulation and technology,” says Piasecki.
[Wall Street's Mission: Cutting Costs While Staying Competitive]
Activ Financial uses specialized field programmable gate arrays, a form of hardware acceleration technology, to process the exponential growth in quotes from equities, options and futures. It now has the capacity of 6 million messages per second, up from 4 million just a short time ago, Piasecki says. “Customers are asking for depth of markets at these levels. It’s getting expensive to interact with the feeds that represent the markets,” he says.
To pick up alpha, firms must see depth-of-market data across all venues. Hence, they’re buying a lot more than the standard feeds from the securities information processors. They’re buying direct feeds to create their own derived National Best Bid and Offers, and they may have dozens of people supporting this expensive infrastructures, he says. Looking back, five years ago having this sort of setup was a unique competitive advantage. Today, the same type of information and feed is essentially an off-the-shelf product with shared economics that can be purchased from a vendor, says Bill Adiletta, a partner in consulting firm Capco’s technology practice. “There’s been an explosion of vendors providing high-performance feed consolidation services and being placed in the colocation facilities,” he says.
A number of vendors offer high-performance ticker plants, including Activ Financial, QuantHouse, Exegy and Xignite. These providers use optimized software techniques and new server designs for scalability, so firms can drive costs down and improve performance, Adiletta says.