A Wrap on NATAS

Every year, my dad, my brother, my cousin and I get together to watch the NFL Draft. Yes, that means sitting around for hours and watching names get called...it's glorious.

As a result, I'm down in Raleigh, North Carolina, with my family, so I'll be slow to respond to emails and phone calls. But before I get back to my vacation, I had a few takeaways from this year's North American Trading Architecture Summit (NATAS).

One of the better discussions at NATAS looked at knowing when to pull the plug on legacy systems. Whether you're a buy-sider or a sell-sider, these are very challenging decisions to make. The key, says Bill Murphy, CTO at Blackstone Group, is communication. From Dan DeFrancesco's story:

"I don't feel like people understand the true cost of not turning it off, and I think we do a really poor job of explaining that as a group," Murphy said during the C-Level Panel. "You don't really understand the anchor it is, the slowness that it's causing in your overall architecture or in your overall organization where people are spending 20 percent of their time on it."

The key, Murphy said, is to put together some type of financial number to show how much an application you're looking to shut down is continuing to cost to run.

The problem is there no exact formula. Murphy is often forced to approximate how much it's costing the firm.

"It's almost a little black magic," Murphy said. "Maybe it's a little bit manipulative in terms of making a best guess, but until they see that number I think they just always assume, oh, it's just no big deal to keep the application going."

The Big Data Conundrum

As always, when talking about infrastructure investment, data management is going to be a concern. This was true for banks and high-frequency trading firms, alike.

Antonio Hallak, CEO at Sibyl Trading, a prop-trading shop focusing on options, lamented that there aren't any technology providers helping in the low-latency space when it comes to big data:

"From a latency perspective, our big focus on a regular basis is taking all this unstructured datasets and looking for opportunities, but also figuring out how to achieve results within a timeline that remains relevant? How do we provide our quants with sentiment from the signals that are relevant for our trading? From a big data perspective, there's no one technology [provider] that has really tackled the high-frequency side, or just the trading side to address the time series aspect of data that comes in and the real-time, or even near-real-time, need we have. There are technologies that are beginning to come out that are relevant, but again the problem with those is that they are very niche."

Meanwhile, in a different panel, Jean-Pascal Chauvet, chief technology officer of equities at Deutsche Bank, said that when it comes to big data, it's more than just innovation, it's about handling the scale of a firm's infrastructure.

"How you deal with this huge amount of fixed cost and running on Windows 2003 servers, and being stuck with 30 datacenters that are these massive things that just cost you money," Chauvet said. "It's very hard to move to newer things. I think one element of not investing into new innovation is that it does not really answer how we solve the problem of dealing with legacy."

But he added that banks better understand what they're looking for when it comes to big data.

"Today, the business is saying, 'now we understand the value of big data', and the IT department is saying, 'okay, we have to also understand how we are going to do this', and therefore they're investing into new technologies," Chauvet said. "There are two or three banks that I see investing into creating labs in California, China, Israel, and elsewhere, but those centers are tech companies that are completely owned by the big bank, and provide the sort of joint-venture mechanism so that we can get into that innovation."

For everything that has been published from the event, and for everything that will be published, you can find it all here.

Five Random Thoughts/Links

* Speaking of the draft, I feel bad for the teams at the top of the draft that need QBs. After watching Jameis Winston and Marcus Mariota in college, I'd be surprised if either of them turns into a franchise QB. If I was forced to pick between the two, give me Mariota. But to be honest, I don't like any of the QBs in this draft. If Bryce Petty was to fall, maybe.

Then again, I'm a Patriots fan, so I haven't had to worry about drafting a QB for a long time...you know, except for when coach Bill Belichick drafts Tom Brady's backup in the second round.

* At this year's Sell-Side Technology Awards dinner, the special guest speaker was former heavyweight boxer Gerry Cooney. He's truly a lovely, fun-loving man, and he definitely entertained the audience.

As a boxing historian (at least I consider myself a boxing historian) I've always felt that Cooney got a raw deal. If you bought a ticket to watch a Cooney fight, win or lose, you were always entertained. Stop thinking of him as Rocky Marciano, and start looking at him as a Micky Ward, James Kirkland, Bobby Chacon or Ray Mancini type fighter, and he was a heavyweight providing those fireworks.

I highly recommend reading the great Gary Smith's profile of Cooney from 1987.

* If you're ever down in Raleigh, there are two breweries that you should check out: Big Boss Brewing Company and Lonerider Brewing Company. Oh, and take the 30-minute trip out to Chapel Hill and get the fried-chicken sandwich at the Sunrise Biscuit Kitchen.

* This year's crop of 3-year-olds running in the Kentucky Derby is SO GOOD. I'm hoping that this race comes down to Dortmund and American Pharoah (yes, that's the correct spelling). I'll be pulling for Dortmund, since he's sired by my all-time favorite horse, Big Brown.

* My prediction for the mega Floyd Mayweather-Manny Pacquiao fight: Mayweather 116-112.

But keep this in mind: For the wet blankets complaining that this fight is five years too late, that may actually help make this a more entertaining, competitive fight. Five years ago, Mayweather would've schooled Pacquiao. But today's Money May is a tad less mobile and the reflexes, while still sterling, are a hair slower...that can only help PacMan, as power is the last thing to go.

My girlfriend is from the Philippines, so I think I'm going to have to root for Pacquiao, although, my brain says that Mayweather will be victorious. But don't discount a draw...boxing loves controversy.