Testing the Waters

As streaming data pours into trading firms at higher rates of speed, new capabilities for position management are following—but more slowly and less visibly.

The growing proliferation of lower latency data feeds is fueling new product offerings in storage systems, analytics, and alert engines for intraday position management and other offshoots of algorithmic trading. As some industry analysts see it, intraday position management technology has really emerged only over the past 12 to 18 months. But early users of the new technology—also known as real-time position management or real-time position keeping—include major names, such as Barclays, but also smaller institutional investors.

Using intraday position management, traders can view, query, and execute transactions on position data from multiple data feeds, either as individual streams or as aggregated data, all just about in real-time. The aggregated data can include positions across multiple types of securities, says Gavin Little-Gill, research area director for investment management and brokerage and wealth management at consultancy TowerGroup.

"If I'm managing a portfolio, I want streaming data—and I want it to feed into my valuation process on an intraday basis. I might write business rules that say, 'If this happens, I want it to automatically trigger an order,' or, 'I want an alert to be sent to the human portfolio manager,'" says Tom Price, senior analyst for securities and capital markets, also at TowerGroup.

Although some traders have been performing position management for quite some time, analysis typically didn't happen until after the end of the day, says Jeff Wootton, vice president of product strategies for Aleri Labs, a company that produces a streaming data storage platform now installed by Barclays and other customers.

"A couple of hedge funds were really the first to use streaming data for real-time position management. Then the big players at the banks jumped in," says Ron Verstappen, CEO of Wombat, the vendor behind a streaming management middleware system currently in use by seven of the world's 10 largest banks. Wombat's infrastructure already interfaces to about 80 different data feeds.

"I'd imagine that all our banks are probably using streaming data for position management," says Wombat COO Danny Moore.

Meanwhile, trading firms are exploiting data streaming as a platform for related solutions, such as risk management, fraud detection, and liquidity management. Barclays first installed Aleri's storage and analysis platform in early 2005, using the product for applications that revolve squarely around position management, Wootton says. In March of this year, Barclays upgraded to the vendor's more scalable, higher capacity second-generation system.

Also over recent months, Wombat has publicly rolled out streaming data infrastructure and integration deals with Bear Stearns, Merrill Lynch and EdgeTrade. With connectivity to Wombat-supported US data feeds just about wrapped up, Bear Stearns is now working with the middleware vendor on integrating data from Europe and Asia. Meanwhile, the wholesale bank has started using Wombat for OPRA, a direct options feed. Bear Stearns has also used Wombat's APIs to build an interface to Vhayu Technologies' Velocity, another streaming data storage system that can be used for intraday position management.

Some smaller banks, hedge funds, and brokerages are taking a more gradual approach to buying and building streaming data solutions. "Slow and steady wins the race," says Chuck Kennedy, managing director for Piper Jaffray, which began implementing the Velocity data storage system about a year ago.

Ultimately, however, streaming data management will become a competitive necessity for all brokerages, says Sang Lee, managing partner at Aite Group, a Boston-based consultancy.

"If you're going to react in as close to real time as you can, you need to know the current state of the market," says Wombat's Moore. "And if you're building an automated trading system, you'll be putting in constraints. You might say, 'Go ahead and trade that—but not if the market goes up or down by a certain amount.'" A company doing intraday position management might set the system to automatically stop trading if the P&L swings too far in either direction, says Moore.

"You might also want to know what it would take to liquidate a position, and you'd need quotes for that," says John Coulter, vice president of marketing and development at Vhayu. For example, Barclays Bank now uses position management for managing its cash positions globally, calculating its intraday positions by geography and currency. Beyond uncovering its global net positions, the bank wants to make sure it has adequate liquidity in regions using specific currencies, says Aleri's Wootton.

Some observers suggest that real-time position management systems can be a tough solution for a smaller firm to put together, due to reasons related to both systems integration and the need for specialized algorithms. Others say that global and local regulations, like Basel II and Reg NMS, will almost certainly act as drivers toward wider deployments of intraday position keeping.

"Reg NMS says, 'If you're providing services for investors, we want you to prove you're giving the best execution. We don't want just a partial look at liquidity. We want to know where all of it is. Tell us where you traded that customer and prove to us that you didn't trade outside where you should have traded,'" says Moore.

"Reg NMS will require a snapshot against executions to ensure compliance," says Vhayu's Coulter.

Piper Jaffray is looking to use Vhayu's Velocity as a compliance tool for Reg NMS. In an implementation that began about a year ago, Piper Jaffray now uses the platform to compare performance of order histories versus industry benchmarks. "We can store and query up to about a year's worth of data in Velocity. This is something that is getting to be critical and still somewhat unique," says Piper Jaffray's Kennedy. Over the next few months, however, Kennedy intends to speak with Piper Jaffray's compliance department about using the platform to query streaming data as an aid in Reg NMS compliance. "For example, we'll be able to ask, 'What are the other bids that are out there for this?'" he says.

Kennedy says that Vhayu's support for Level 1 and Level 2 data is important. "It gives us access to both trade and quote data," he says.

Also, within the next few months, Piper Jaffray expects to start using the Vhayu platform for the intraday playback of certain events, such as exceptionally large transactions by other traders. Storage platforms from Vhayu and other vendors constitute a set of streaming data facilitators, says Aite Group's Lee. Other vendors in this general category include Aleri Labs, Datastream, iSpheres, Kx Systems, Progress Software and Xenomorph. Some of these storage systems support ticker data only, whereas others also process real-time and historical data, Lee says.

In another subcategory, products such as Progress Software's Apama deal with ticker and transaction data as events, says Lee. Many systems process data in memory only. Kx Systems, on the other hand, is more of a database play. "Kx's product inputs data into sort of a relational database," Lee says. Streaming data storage systems typically come with either pre-built analytics or programming tools for creating analytics. If not, they are able to interface to outside analytics or programming tools.

What kinds of queries might be raised about streaming data in intraday position management? "You might ask, for example, about your total exposure to the automotive market," says TowerGroup's Little-Gill.

Thomas Weisel Partners (TWP), another Vhayu customer, stepped to the Vhayu platform from a competing product in August of 2004. According to Dhiru Patel, TWP's chief quantitative strategist, Vhayu's platform has allowed TWP to move into high-speed quantitative trading while at the same time continuing to offer services that compare performance of customers' portfolios against benchmarks. "The entire Street was moving toward low-touch trading. We wanted to capture this business so it wouldn't go elsewhere. We also wanted to increase our product line so we could generate more order flow," Patel says. Unlike Piper-Jaffray, TWP has no plans to use the technology for compliance with Reg NMS. Large portions of TWP's back-end processing are handled by the trader's OMS providers, who will also help take care of TWP's Reg NMS compliance.

TWP selected Velocity because of its nanosecond performance, as well as a C++-enabled API for quick customization and integration with internal and external systems. TWP uses some of the platform's built-in analytics, says Patel, but the firm has also embedded its own custom C++ code in Velocity's analytics engine to answer special requests from customers for historical data analysis.

In contrast, some other products still force firms to use proprietary programming languages. Yet developers who know these proprietary languages can be difficult to locate and hire. Along the way, Vhayu has also constructed an interface to integrate its system with TWP's Stream data feed. "It's also in our plan to integrate options data and futures data, although I can't tell you how much or when," says Patel.

Trading firms are generally reluctant to talk about exactly how streaming data will come into play for intraday position management and related applications. Some observers attribute the silence about specifics to competitive rivalries among traders. These days, trading firms employ mathematical wizards who specialize in writing algorithms for quick analysis of streaming data, says Simon Garland, CTO at Kx Systems. "Somebody can come up with an algorithm in the shower one morning, and later, make millions of dollars just from that single algorithm," he says.

Known in some circles as financial engineers, these algorithm gurus are either traders themselves, or work closely with them, says Garland. Companies like to keep the analytical secrets gleaned from these wizards—and the technology that supports the traders behind the scenes—very close to their vests. Still, Kx Systems counts Merrill Lynch, JPMorgan, Lehman Brothers, and Deutsche Bank among its customers.

At the same time, with more trading firms looking into streaming data management, competition among vendors is heating up. However, intraday position management entails more than just abundant data feeds and effective analytical tools, say industry watchers. Other ingredients can include some kind of alert and exception management mechanism, plus integration with back-end portfolio valuation systems and OMSes. Increasing numbers of off-the-shelf products are now becoming available in these categories. "But there's a ton of proprietary build out there, too," says TowerGroup's Price.

For their part, vendors grapple with issues around what kinds of features to add for streaming data customers, as well as how to integrate their products with disparate trading systems. Beyond performing data analytics for customers, Xenomorph is also testing new off-the-shelf analytics. Xenomorph does not provide links to OMSes, say officials from the data management software vendor, but some Xenomorph customers have been funneling the results of pre-trade analysis into OMSes from Royalblue and GL Trade.

Meanwhile, some customers have used Wombat's tools to build their own alert engines, says Verstappen. Within the next few months, however, Wombat plans to release an off-the-shelf alert system, he says. The company is also thinking about readying a pre-built order management system . "But that will depend on customer demand," he says.

Wombat's recent announcement with Merrill Lynch seems to represent a particularly ambitious systems integration outsourcing effort, says TowerGroup's Price. "But the Merrill Lynch deal really appears to be a bit of an anomaly. I'm not sure whether other firms will be willing to rip out and replace to the same extent," he says.

What's much more certain is that traders intrigued by intraday position management now have a wide selection of products to choose from. "This is a not a bad thing. It speaks to the narrowness of specific solutions, but also to the complexity of trading environments. Sometimes the products are substitutes, and sometimes not," says Price. For example, a bank might decide to use products from multiple vendors for storing and analyzing streaming data. "Each of the streaming data products out there has its own pros and cons. These products are kind of like kids. You can love them all, but for different reasons," Price says.