Data-as-a-Service Emerges As Valuable Concept

Blog entry

The data management method known as Data-as-a-Service (DaaS) can be very effective for addressing data operations challenges that have emerged due to regulation and for other reasons, according to data and analytics executives who spoke in a webinar hosted by Intelligent Trading Technology on June 15.

Webinar Recording: Data as a Service: Realizing its Value for Data Management

Increased volumes of market and reference data are one reason for interest in DaaS, according to Thomas Kennedy, head of analytic services at Thomson Reuters, which operates the Velocity Analytics DaaS platform. “To be able to annotate that information and properly analyse it, we’ve seen firms strategically consolidate the cross-asset challenge by taking a platform-based approach,” he said.

As a platform, the DaaS method can leverage web-scale technologies to let firms manage market data and order and execution data more effectively, and also generate analytics to support trade surveillance, best execution and transaction cost analysis.

MiFID II best execution requirements entail comparing numerous liquidity pools, which requires a “huge amount of data and analysis,” said James Corcoran, head of engineering at Kx Systems, which provides analytics to power the Velocity Analytics platform. “If you’re looking to build a best execution analysis, DaaS lends itself well to dealing with the number of different analytics required for best execution support, and how those analytics vary across asset classes.

Featured Download: Poll results on Data as a Service: Realizing its Value for Data Management from our recent webinar audience

Aside from pressing regulatory concerns including MiFID II, DaaS addresses the “big data problems” firms have, added Corcoran. “For the first time, firms really have to aggregate, cleanse and normalise data from a huge variety of systems and disparate data sources,” he said. “Having a DaaS platform in your organisation eases some of that pain, even for reporting a trade across various different asset classes. Most buy- and sell-side firms have a range of different trading applications and systems where data needs to be collected and aggregated.”

Analytics Benefits

Tyler Capital, a London-based proprietary trading firm, has been able to make its data operations more flexible and manageable with DaaS from Thomson Reuters, according to John Denheen, head of data at the firm. “If we want to expand and move into other business areas, we have to record the data first before we can back-test,” he said. “If we want to start trading a new product and need three years of history, we don’t have three years to wait to start trading.”

Tyler Capital can acquire the data it needs, perform analytics and prepare the data to evaluate volumes, market behavior and volatility, Denheen explained. “Rather than taking time coming up with analytics for data in that shape, it is better to just move our analytics to that [DaaS] platform in the cloud and to those broad stroke analyses,” he said. “The possibilities being opened up by this new way of using DaaS are exciting.”

Faster access to data is the key to a successful DaaS effort, according to Kennedy of Thomson Reuters. “We looked at how to bring the technology closer together, how we provide more efficient normalisation, and out-of-the-box usage of content whether it’s streaming information on our streaming network, historical data sets, reference data fundamentals or news,” he said. “People want access to that to build analytics.”

Cloud Advantages

In recent years, Tyler Capital became more willing to use cloud computing to back-test data, according to Denheen, even though “you still have an environment where you move your data to the cloud, or data has to be present in the cloud – or you have cross-connections with your infrastructure – which has its own risks,” he said.

Tyler Capital co-locates its trading operations, and chooses not to move that to the cloud, Denheen added. “It’s an effort to store and manage market data efficiently,” he said. “If some of that workload can be shared, it makes our job of running our institution a lot easier.”

Cloud computing is best suited to support DaaS in two areas, according to Corcoran – capturing data all day, every day, with back-up and recovery capability; and “doing bursts of computation, where you want temporary or transient workloads,” he said.

To optimise data handling performance, however, which requires flexible operations in a public cloud environment, requires paying a premium, noted Kennedy. The evolution of DaaS can improve upon the cost economics for that function, he said.