Combining compression with a tick database should help cut storage costs for Wall Street firms, Hudson asserts. "Today people get charged so much per terabyte for disk usage — anywhere from $20,000 to $50,000 per year," he says. "So someone storing 20 TB at $20,000 a year is paying $400,000 annually. By using Squeezer, their cost for storage goes down to $100,000 per year, because they only need a quarter of the space. The other way some of our customers look at this is they can grow their data four times without buying more disk [space]."
Some Wall Street executives interested in using FPGAs to accelerate applications, or portions of them, however, have expressed the concern that it's hard to find programmers who are skilled at writing applications for FPGAs. "It absolutely is," concedes Hudson. "It's tricky because the software and hardware have to work hand in glove together. This isn't like installing a new network card." Hudson notes that Vhayu, based in Silicon Valley, has been in a good position to hire FPGA-savvy programmers.
Vhayu engineers, Hudson adds, designed the FPGA card in Squeezer specifically to optimize it for market data. "There were a lot of interesting considerations about the type of data, how the data is represented and how to make this work," Hudson explains. "The real challenge is in making sure it can maintain enterprise scalability with tolerance and reliability, because as the amount of data gets bigger and moves faster, if something breaks or fails, there's no time to catch up. At the end of the day, this stuff can't break."