SIFMA Roundtable: Can Tech Turn the Tide?

The SIFMA technology conference has become an annual pilgrimage for financial technologists to seek out the latest developments and data sources, and to debate with their counterparts and suppliers. But in recent years the financial crisis, a troubled economy and tightening budgets at user firms has meant fewer exhibitors. However, this has not dampened the pace of development, with data and software vendors continuing to roll out new advances to address the most pressing needs of the industry. Here, Inside Market Data assembles a panel of experts from a range of vendors to discuss the hot issues at SIFMA 2010

IMD: What will be the hot topics and innovative technologies at SIFMA this year?

John Bates, chief technology officer, Progress Software: Let's face it, the hot topics of this year are regulation and the "flash crash," so technologies that help firms appease regulators with things like real-time risk management and real-time market surveillance will be hot. Technologies that can detect patterns that indicate a potential flash crash is happening and either initiate circuit breakers or take some real-time remedial action are also cool.
And the high-frequency trading war is always hot: It's the equivalent of sports cars competing on horsepower and handling, and technologies that enable lower-latency trading and more rapid strategy evolution-whether at the messaging level (e.g. special hardware or new techniques), the data level (e.g. amazing new tick databases) or the platform level (e.g. new enhancements in complex event processing)-are always winners.

Richard Chmiel, vice president of global sales, OneMarketData: We believe that CEP technology will continue to be at the forefront of SIFMA, since it is a very convenient paradigm for such a large class of applications. Given that a lot of decisions in finance are event-driven-where events range from highly structured real-time streaming market data to completely unstructured events such as news articles-many financial applications are natural candidates for CEP. This includes not only the obvious candidates such as high-frequency trading, surveillance and compliance, order book aggregation, and smart order-routing, but also risk management, P&L and exposure calculations, and much more.
With CEP, users can conveniently decipher, normalize, analyze, compute and publish real-time trading signals, perform real-time book building, investigate and act on inherent liquidity, and easily implement a myriad of programmatic statistical arbitrage trading strategies.

Laura Perdue, vice president of marketing, Chi-X Global Technology: While everything with the label "low latency" was hot last year, this year the industry is coming back to the importance of functionality as well. Another hot topic is latency measurement as a competitive tool.

Jeff Banker, executive vice president of Real-Time Market Data and Trading Solutions at Interactive Data: Latency management is top-of-mind for many of the types of firms at the SIFMA show this year. Financial institutions are increasingly attempting to improve latency, but they need to be equipped with the right tools for measurement and monitoring.
In addition, there is a lot of discussion around cloud computing and Software-as-a-Service (SaaS). Firms are looking to minimize costs, speed deployment, reduce bandwidth requirements and quickly access enhanced services. Interactive Data is one of the largest providers of hosted market data infrastructure and applications to meet these needs.

Neil McGovern, director of marketing, Sybase: Real-time analytics will continue to be a key technology-we are seeing an uptick in interest in real-time liquidity management outside of North America in particular. Innovative solutions that can help generate an up-to-date, enterprise-wide picture of risk, positions, etc. will also be important.

Martin Hakker, executive vice president of marketing, Fidessa: Naturally, there will be a diverse mix of new technologies on display, but it would be reasonable to expect the hot topics and the innovative technologies to overlap, as successful solutions are those developed in response to market drivers such as the steady increase in electronic trading. Given the current expectation of a new and more stringent regulatory climate, solutions that ensure compliance with current and upcoming rules and compliance guidelines will be fundamental, particularly those in the risk management arena.

IMD: What technology-sensitive parts of their market data sourcing and distribution infrastructure are firms looking to address right now?

Gordon Bloor, chief executive, Morningstar Real-Time Data: Interoperability: normalization and mapping between different sources across distribution infrastructures is back on the agenda-not just as a technology issue, but encompassing commercial restrictions, licensing, and, hopefully soon, regulatory direction from the EU. It is still one of the biggest barriers to genuine choice in the data distribution space.

Hakker: For many organizations, simply managing the sheer increase in market data volume presents a challenge in itself. That said, the drive to reduce latency is at the heart of most firms' strategies. At the same time, companies will be looking to increase throughput and guarantee best execution as market data volumes continue to skyrocket and certain markets remain volatile.

Banker: My observation is that firms are seeking technology solutions that can assist in managing their expenses through infrastructure optimization. Also, while firms that are not latency sensitive are trying to manage the increasing costs of market data, other firms are focused solely on achieving latency objectives. They are attempting to manage costs while optimizing latency. Solutions that can reduce latency will certainly be in the spotlight at the show.

Bates: Low latency is always a requirement for high-frequency trading. Many firms have extensive market data infrastructures but are finding latency unsatisfactory, so they are looking at strategies to move to lower-latency solutions. As markets continue to fragment, there are more markets, with more data to be handled. Also, more and more firms want tick databases to store and backtest the data-what's the right strategy? And from a Progress Apama point of view, we've seen growth in interest in CEP as a technology to analyze data at line speeds and respond immediately.

Chmiel: Firms are looking to address two different areas regarding market data. First, those involved in low-latency trading are looking for direct exchange feeds and CEP solutions that allow them to capitalize on latency advantages down to the microsecond level. Yet ultra-high-frequency trading is not the only area where financial institutions are looking to address their market data infrastructures. For example, less latency-sensitive functions, such as risk management departments, depend on real-time market data, yet may be more interested in data capture and analysis via a high-performance database solution than an ultra-low-latency CEP solution.

IMD: What factors are currently driving firms' technology spending decisions and deployments, and how are vendors and data providers responding to these needs?

Perdue: We're seeing more client focus on the issue of control, to maintain competitive edge in a fast-moving market landscape. Exchanges and participants want the advantages of a vendor solution (managed, supported, proven) combined with flexibility in terms of deployment and the business relationship. Seeing this trend early on, we adopted a development and business philosophy to support it from the beginning.

Bloor: In light of many of the regulatory concerns and the Securities and Exchange Commission's investigations into market activity, one factor having a growing influence on technology direction is compliance. An example of this is the growing demand for detailed post-trade analysis driving demand for full market-depth tick data.
Data delivery performance-not just in terms of latency but also the continued dramatic increase in volumes-is also driving spend. Some clients or vendors will continually throw more budget at communications and infrastructure, but increasingly firms are looking for innovative delivery for individual applications, such as bespoke filtering and netting or the pre-processing of data, all of which can be used to reduce data collection overheads.

Hakker: Many firms are making their technology investments with an eye to reducing total cost of ownership. Disparate systems can create technology headaches and be costly to solve, while consolidation among vendors and end-users creates new challenges every day. Technology decision-makers must remember to resist market hype and doomsday scenarios and make purchases that will yield strong returns and set them up to adapt future technologies and additional systems easily and cost-effectively.

Chmiel: Today's firms are looking for solutions that help them keep costs low and increase efficiencies across the market data and trading workflows. They are also looking for solutions that will give them a competitive edge and help increase profitability, whether it be reducing market data latency or eliminating human errors associated with porting code among multiple systems. Currently, vendors and data providers are responding to these needs by offering more trials and allowing firms to determine for themselves whether a specific solution will provide an adequate return on investment. This environment also forces vendors to examine the flexibility of their solutions. Those solutions that benefit multiple departments and can be deployed globally offer firms true value that in-house or narrowly focused vendor systems cannot replicate.

Bates: There are three main themes driving our business: Managing pre-trade risk and monitoring, real-time market monitoring and surveillance, and FX liquidity aggregation. For example, pre-trade risk and monitoring for sponsored access in real-time-clearers that offer sponsored access need to ensure that the firms using their exchange memberships are not doing crazy things. I believe firms will develop solutions in this space. In the surveillance space, following raids by regulators, large firms are very conscious that they need to show they are monitoring unusual market occurrences, such as large price and volume movements in all instruments, as well as potential internal breaches of regulation, and surveillance providers are trying to focus more on real-time services in response. Finally, around the world we see continued interest among trading firms in trading FX from one screen that shows an aggregated view of ECNs and bank portals. This can be used on the spot desk, for auto-trading and for custom FX pricing. The benefits are radically increased profits through more intelligent operations, and vendors are developing solutions here-as are the trading venues and banks.

McGovern: Our customers and prospects are very sensitive to vendor viability and time to market. Small, high-return projects with safe vendors are more popular than large projects, and vendors such as Sybase are working on programs that not only sell innovative technology, but provide "rapid-start" service offerings to get projects moving more quickly.

Banker: Latency-sensitive trading is being more widely recognized as an asset, and direct strategies are increasingly being adopted and implemented. Financial firms are also looking for ways to control or even eliminate additional technology spend on servers, circuits and routers that have traditionally been needed to support market data applications in a deployed environment, prompting them to review cloud computing as an alternative option for supporting their business processes. Cloud computing can also help them realize significant reductions in bandwidth requirements, as a cloud solution can provide only the levels of data that they require, instead of all the data their vendor wants to deliver to them.

IMD: We've seen increased interest in latency monitoring technologies. Is this something that will remain the preserve of niche providers, or will mainstream vendors seek to incorporate this into their offering?

Perdue: We agree-there is tremendous interest. We are incorporating it into our MarketPrizm managed trading infrastructure service.

Banker: We are beginning to see vendors incorporate latency tools into their offerings, usually by partnering with smaller, more specialized providers. At Interactive Data, we incorporate the capabilities of a number of vendors, in addition to our own developments, to provide customers the latency metrics they require. This can allow customers to benefit from the competencies of niche providers, as well as the high performance and low total cost of ownership of our managed infrastructure services.

Chmiel: With more firms competing for profits in the low-latency space, there is an increasing need for low-latency monitoring technologies to support troubleshooting of specific orders and determine where latency enters specific workflows. However, new Federal regulations, along with an overcrowding of firms within the ultra-high-frequency space, may also lead traders to seek out alternative strategies that are less focused on speed and more focused on unique ideas. As such, it is too early to say whether mainstream vendors will begin incorporating this technology into their offerings.
Bates: We see customers wanting more and more capabilities to analyze where the time is going in a solution and whether there are opportunities to squeeze any of that latency out. And I think this goes for every part of the trading value chain. That's why we have built advanced profiling and latency measuring tools into Apama.

McGovern: While there will continue to be a market for "pure-play" low-latency technology, at Sybase we are convinced that low-latency solutions are only one part of the whole "real-time analytics" stack. As more vendors add low-latency technology (either through in-house development or OEM deals), we expect the balance to tip very heavily against "pure-play" in the next few years.
I do think latency monitoring will remain a niche technology. There's a limited market for it, so it fits well with specialist providers. There's an interesting question, however, of how the latency monitoring technology is used, in terms of what you do with the information it produces, and we've seen interest in taking the latency data as in input into a Sybase CEP engine to help decide what actions to take based on the data.

Bloor: By default if you are selling purely on a latency play, benchmarking is critical, and more vendors will have to deliver effective measurement and monitoring technologies as part of their core offerings. For a wider audience than HFT players, and across the high-performance global consolidated feeds, the completeness and accuracy of the data, providing full depth and every message available, is of greater importance.

IMD: Has the SEC's investigation of high-frequency trading worried providers of low-latency data and technology who serve that marketplace, and are they now finding new ways to use their technologies in other business areas?

Bates: It depends on the vendors. I think the trading firms are worried. Vendors with inflexible legacy products are worried. But some vendors are welcoming the SEC and the Commodity Futures Trading Commission looking into algorithmic trading, the flash crash and what to do. For example, Progress Apama has long used its CEP platform to empower real-time risk, monitoring and surveillance applications. In fact, these are used by regulators around the world, such as the UK's Financial Services Authority. I think that providers of low-latency data and other technologies are well positioned in a world where regulators police the markets, as the regulators themselves-like the trading firms-will need access. However, if the regulators were to ban or restrict high-frequency trading, that might be cause for concern to some.

McGovern: I don't think the investigation could have cheered the industry up. Reading the transcript of the SEC roundtable must cause concern that the SEC is going to look deeper into the issue, particularly around proprietary information, co-location and the possibility of front running (which was raised at the roundtable). I think the concerns with HFT are misunderstood by the mainstream media. And, the fact that this singular label is being loosely applied to a range of trading practices-some of which may increase market volatility, while others are generally recognized as improving market liquidity-adds to the confusion. The markets are very agile and adapt to changing market structure and changing technology. Even if we see the market backing off some of the current HFT practices, it won't reduce the demand for low-latency market data and real-time analytics to understand the markets, make decisions on trades, and monitor exposure to the markets and to creditors.

Bloor: The SEC's reviews are wide-ranging and could easily lead to more regulation that will impact HFT firms. The knock-on effect for those firms focused purely on technology to support this activity could well be significant, and while few are openly discussing their concerns, I'm sure they are all actively reviewing opportunities to widen use of their technologies.

Banker: While there is a segment of the marketplace that is highly exposed to potential changes in regulation, I think the broader marketplace still has the same goal in mind when it comes to trading: They seek the fastest paths across asset classes and exchanges, and they want to deploy various trading strategies that either benefit from, or have lower risks than, not being at a latency disadvantage.

Chmiel: As of now, we have not seen any difference in demand for low-latency data or technologies from our clients. However, should new regulations change the low-latency workflow, there will still be a large need for market data and technology solutions, including CEP, whether for intra-day quantitative trading, risk monitoring or a number of other functions. Financial firms will always need tick data as well as the technologies that can capture, store, process and analyze it.

Perdue: We're all watching this space. But whatever develops on the regulatory front, remember that "high frequency" is just one electronic trading strategy that depends on technology infrastructure with speed, reliability, scalability and flexibility. As a technology provider, our work is developing products that support our clients' businesses as the competitive and regulatory environments continue to inevitably evolve.

IMD: CEP is an example of one technology that has been expanding into risk, liquidity management and other areas beyond its base in algorithmic trading. However, the CEP industry has also been consolidating. What does this mean for CEP and similar parts of the industry?

McGovern: This is good news for customers. The consolidation shows that vendors such as Sybase understand that a commitment to CEP is vital for real-time analytics-and the recent moves by Sybase show our commitment to be number one in the CEP market for real-time analytics needs, such as trading, product development, risk and liquidity.

Bates: I've said for a long time that small CEP startups just can't survive in the more mature CEP market. And the economy over the last two years accelerated that for some vendors like Aleri. There's a worldwide infrastructure of sales channels, geographic coverage, marketing, support (to name but a few aspects) that are needed to properly give confidence to global customers that they are being properly supported. Ideally you need the innovation of a startup and the backing of a big company-that's certainly our goal at Progress.
Most major companies see the power of CEP to embed into certain solutions. But this doesn't mean they're all competitive with each other. I also believe that CEP alone is not always enough to build the full solution. We've seen the benefit of combining CEP with Business Process Management (BPM) to manage human workflow and system orchestration in, for example, a market surveillance solution.

Chmiel: CEP began as a stand-alone technology focused exclusively on low-latency trading, and in recent years has emerged as one of the leading new technologies at financial institutions. Yet leveraging and deploying multiple disparate systems for algorithmic trading, risk management and more is time consuming, expensive and can introduce human errors into the workflow. Consolidation of CEP with more traditional data management vendors is a natural progression as CEP enters the mainstream.
We believe vendors across the market data and data management fields will continue to consolidate to form integrated data management systems where technologies such as CEP become standard components of larger, unified platforms. However, the pitfall remains that unless these systems are built from inception for these uses, many of the problems originating from disparate solutions will remain. Thus, the true winners will be those vendors with the foresight to build their systems natively to include all these functionalities.

Bloor: CEP technology is becoming more mainstream. Consolidation among the CEP players will allow the fundamental technology to be incorporated in enterprise solutions and better integrated with other tools, thus widening its use to different business areas and industries. As the CEP technology spreads across applications and the enterprise, it will be interesting to see how it both interacts and conflicts with low-latency providers. Consolidation of vendors across these two spaces would be logical but has yet to appear.