AFTER HAVING ATTENDED the Securities Industry and Financial Markets Association’s (SIFMA) annual technology conference in Manhattan for several years, this years event came as quite a shock. Attendance was down noticeably, and several big names were not in attendance.
Since SIFMA is the conference for financial services technology, it’s a great place to see and be seen. In years past you might also accidentally discover something new and useful while you were there, that is the raison d’tere of a trade show. This year, the financial services sector has been particularly hard hit by the recession, with huge companies like Lehman Brothers and Bear Stearns simply going away. Many other other companies in the sector are under severe financial duress.
You would normally think that the rotten economy would spur innovation for the businesses that serve the financial services sector, but sadly, that was not the case. This year the conference showcased the same stale technologies as in previous years, only less of it. The diminished size of this years show reflected the degree that economic conditions have decimated the financial technology industry. Normally there are three floors packed full of exhibits, this year there were only two as prominent exhibitors of years past such as Reuters, Microsoft and Sun were nowhere to be found.
The show was pretty bleak overall, but there were a few noteworthy things to see in spite of diminished size. For example, the lunch presentation put on by NYSE Technologies to showcase their Version 5 Financial Data Platform launch was disheartening to say the least. On the other side of things, Intel Senior Fellow Steve Palowski spoke about how multi-core architectures can help market data firms cope with increasing processor demands.
These demands are driven by exponentially increasing trade and quote message rates that are a reflection of increased activity in the algorithmic trading space. Intel loves these guys. In addition to message rate demands, there is an latency arms race between algorithmic trading firms. Intel also likes this. Basically, if you can do what you do faster than your competitor does, you win in the algo space, meaning potentially billions of dollars.
Message rates have been increasing faster than Moore’s Law for the past eight or nine years. Mr. Palowski was describing future advances in multi-core architecture while touting these advances as the solution to the problems keeping up with trading and latency demands. He said, “Intel views the financial services industry as kind of a canary in the coal mine because the demands are typically three or four years ahead of other industries”.
Unfortunately he didn’t really address the real problem which is not hardware related but has more to do with the complexity of the software required to take best advantage of these multi-core architectures. There just aren’t many software engineers around who can design and build the required massively multi-threaded software. Because of this, bigger, faster and more cores in the hardware department really buys nothing except a higher power bill and more heat. Intel doesn’t like it when you say things like this.
We were expecting to see more vendors at SIFMA touting cloud computing since that seems to be the buzzword of the day. We really didn’t see much of that though. One thing that was quite noticeable is that the proprietary software vendors who are selling general purpose products such as database software are becoming very scarce. Financial firms are trending towards open source alternatives in a big way.
Microsoft still holds sway in the trading room but seems to be dead in the data center. People we spoke with who were doing back office development were all using various flavors of Linux, proprietary software is dying in the financial space. The exceptions to this rule are centered around specialized products like trading station software and legacy development. If your developers’ expertise is centered around Microsoft products or you are developing desktop software, the chances are you’ll stay with proprietary software. Everyone else is using open source for their new development, maybe that was part of the reason that there were so few booths this year.S|A
Latest posts by John (see all)
- Wall Street tech conferences are not immune to downturns - Jul 1, 2009