BST Awards 2012: Winners' Circle─Progress Software

What are the biggest challenges facing your customers in the market at the moment, and what do you see being important in the immediate future?
Balancing the need to invest in technology, be it for regulatory or business reasons, at a time when IT spend is heavily constrained. In risk and compliance, for example, we see customers addressing pretty well-understood requirements, like surveying for market abuse, but having to rapidly apply those techniques to new markets like energy or swap execution facilities (SEFs). Equally, new requirements such as monitoring automated trading for erroneous orders that might cause mayhem suddenly become a priority.

The sheer quantity of new requirements and the rate at which technology needs to adapt is genuinely daunting. Even though motivated by increased regulatory and media scrutiny, the investment is simply too great for many firms.

As well as its standard use-cases, in what ways can CEP be utilized across the enterprise?
CEP is a great technology when actions need to be taken based on rules being applied to high-velocity, high-volume data flows. The image that springs to mind is CEP as a niche technology for algo trading and related areas, but that image is somewhat out of date and based on the early adoption of the technology. It doesn't accurately capture the current reality-that CEP is mainstream technology used across capital markets.

Today, our customers look to CEP to support numerous time-critical decisions. In the front office, price aggregation of foreign exchange (FX) rates is well established. But if SEFs shape up as expected, there will be a new and urgent need for similar capabilities in what will be a highly fragmented marketplace-capabilities that can be cost-effectively delivered using CEP.

In risk and compliance, CEP is an established solution for firms looking to achieve high performance yet flexible market surveillance. A CEP-based monitoring platform, if implemented correctly, efficiently gathers information from across the business and detects interesting patterns of behavior. The added value of CEP is to extend this capability well beyond detecting insider trading and into fraud monitoring, rogue-trader monitoring, anti-money laundering (AML), pre-trade risk and more─all from a single investment and hence a greatly reduced total cost of ownership (TCO).

We believe that high-frequency trading has peaked, not because of the apparent EU witch-hunt, but because the fundamentals of the business have changed. Performance is now so high and the investment required for incremental gains so great that the pace of change has slowed.

What does Progress Apama have planned for the next year?
Firstly, we are looking into performance. Apama already has excellent low-latency performance and we are looking into ways that Apama can extend its position as the performance CEP product of choice.

Secondly, we are continuing to invest in the Capital Markets Platform and the added-value components that leverage Apama CEP. In areas such as risk management and analytics, we are adding pre-built and supported components to speed the development of new Capital Markets applications, as well as new design tools to enhance productivity.

Thirdly, we want to improve the Apama testing cycle. We want to help our customers mitigate the risk of rolling out new applications with strong back-testing and simulation capabilities.

What, in your view, will be the big technology drivers to watch in 2013?
We believe that high-frequency trading has peaked, not because of the apparent EU witch-hunt, but because the fundamentals of the business have changed. Performance is now so high and the investment required for incremental gains so great that the pace of change has slowed.
Furthermore, volumes are down so the rewards are thinner and thinner. The arms race was simply too expensive.

We also see a shift in emphasis from performance to intelligence where intelligent algos make better use of data to analyze market conditions. This necessarily means more latency as the algos ramp-up complexity but that is the way we see the market going. However, this evolution raises an important issue-the reliability of the algorithms and the need to minimize rogue algos causing mayhem in the marketplace.

We see, as an extension of this shift to more intelligent algos, a rapid growth in CEP-based algorithmic-testing tools that make use of historical and simulated data to push algos further than they are typically pushed today and to drive reliability into a market segment that has, let's face it, had a pretty reputation-damaging year.