Chicago’s Sun Trading Solutions is expanding its use of the Kx Systems kdb+ database across geographies and asset types.
“One reason Sun is using us is they are capturing new asset classes and new geographies and they wanted to go with one solution,” said Fintan Quill, senior engineer at Kx Systems. Some firms have separate systems for different asset classes, such as foreign exchange, equities, and options. For cross-asset trading, said Quill, it makes sense to have the same technology. Kx has been working with Sun Trading since 2008, he added.
“As a long-time user of the Kx database, we have been consistently pleased with how the real-time analytics and speed of the platform has allowed us to quickly bring trading ideas to the market,” said Kurt Lingel, chief technology officer at Sun Trading. “We will use the Kx technology to address our growing research and analysis needs as we continue to diversify across multiple asset classes and markets.”
Kx handles huge volumes of market data, including the spikes that some other systems have to discard, with a highly efficient program - its binary executable file is just over 200 kilobytes. Kx uses a single architecture for both real-time and historical data.
Quill said some firms are turning to cross-asset trading because the high frequency trading (HFT) business is now dominated by ultra high frequency trading firms whose massive investments in sophisticated hardware like FPGAs has made it difficult for smaller players to compete.
“Margins are becoming thinner and thinner and the number of players smaller and smaller.” He sees a move to “quantamental” trading, a combination of quantitative and fundamental analysis. Some HFT firms have found they are spending so much on fancy, and expensive, hardware that they aren’t making as much as they had expected from trading.
Like several other industry experts who have commented recently, Quill said that IT budgets are tighter as volumes and revenues decline. Firms are looking at ways to manage their infrastructures better.
“If you can get a technology that will manage all these datasets on one platform, you are ahead in the number of people you have to employ, and your time to market is faster . One advantage of kdb+ is you are working directly on the data, not taking it out of a traditional database and massaging it. You can build your analytics directly on the data, so you can put your strategies together fast, and test fast and get them into the market.”
Testing can become a significant delay in implementing trading strategies -- see my post below about the results of a MathWorks survey which found that in some banks it can take months to implement a new algo. A survey by the Federal Reserve Bank in Chicago found that a surprising number of trading firms were deploying modified algos without testing them.
Quill said Kx offers a very robust backtesting environment.
“With Kx, you can test very very fast and put your algos into the market.”
