Slashdot Mirror
Wall Street Becoming a Linux Stronghold
alphadogg recommends an article about the rise of Linux on Wall Street. We discussed the beginnings of this trend last year. From NetworkWorld:
"Wall Street firms increasingly are buying into Linux, but some still need convincing that open source licensing and support models won't make using the technology more trouble than it's worth. Linux providers, speaking this week at the Securities Industry and Financial Markets Association conference in New York City, stated their cases that Wall Street firms have nothing to fear about diving into open source. Red Hat and Novell argued that's especially true now that specialized Real Time Linux has been developed that meets strict low-latency and messaging requirements of brokerages and trading firms."
I have the same problem at my work. I want to automate and speed up a lot of the reporting my coworkers do by moving the processing over to one of our Linux servers, but Excel is always a problem. Some of our people actually see Excel as a platform in itself. It's become kind of a joke among some of us there. "Excel would make a great Operating System if only it had a decent spreadsheet."
Some of our macros can take upwards of twenty minutes to run.
I suppose they could use OpenOffice-server, and I've considered playing around with it, but it seems like too much unnecessary overhead. Right now I think I'm gonna give JExcelAPI a whirl as soon as I get a break in between projects.
And Linux will never replace mainframes. Nothing will.
At the risk of being modded troll, OO Calc will probably never replace Excel - other than Suns and big iron, corporate america runs on Microsoft Excel (not necessarily a good thing, but still).
OTOH, I know companies that are still running their websites and outward-facing interface systems on hardware and software that could be easily replaced by off-the shelf open source stuff, which will probably save them a lot of money.
The twitter monologues. Click on my homepage and be amazed.
The kind of use they're talking about isn't really about quants and their modeling. It's about transactional throughput, enterprise messaging, and the guaranteed delivery of various business events (along with the relevant data) to a wide variety of systems across front, mid, and back office domains within a very constrained time window.
As for quants, they often like Linux for a completely seperate reason, specifically because they can use it for Shadow IT purposes without the IT department getting all pissy. Also, many of their favored math packages are old school C and they learned to use them in school on Linux so they tend to gravitate toward it in work as well.
At least that's what I've seen over the last 10-20 years or so since quants have become all the rage.
The article includes a lot of confusion and/or FUD about licensing.
Someone needs to sit down with some of these people and explain to them what the GPL actually says. It doesn't require software written to run on Linux to be GPL'd. Even if you had some reason why you wanted to modify the Linux kernel itself (and why the hell would a Wall Street firm want to!?), you wouldn't need to GPL your modifications unless you were turning around and selling or distributing the modified version publicly.
I work in one of the top 5 Wall Street Firms. Linux is our default OS and represents about 85% of our server deployments. I can tell you that we absolutely do contribute kernel modifications back to the community - the main reason being that when we find kernel bugs (and we do) we need them integrated back into a vendor supported kernel before we'll even consider deploying them into production.I used to have a position where I met quarterly with most of the major Wall Street CTOs/CIOs. Every one of them was heavily involved in deploying Linux. You could sum up their reasons quite simply: commoditization yields cheaper computing.
All of them were tired of being locked into the hardware that Solaris required (i.e., Sun's vertical stack), and paying Veritas Foundation Suite licensing on top of that. (I mean, come on, no big enterprise shop ever used Solaris Disk Suite as a standard!)
Sure, today you can run Solaris on x86 more credibly and there's ZFS, but three years ago that was still vapor. Sun was too late with them.
The writing on the wall for Sun's big servers has been there for some time. Sun could not afford to cannibalize its enterprise offerings by going whole-hog into Solaris x86, which is why it's always been the poor stepchild. In the meantime, Linux came along, reached maturity, and now anyone wanting to buy a Unixy system can let Dell, HP, IBM, Sun, etc. compete to deliver a cheap x86 box. There's no important differentiation between them, and very few people are buying giant Sun servers any more. Heck, Sun's big Lonestar supercomputer sale was commodity x86 running Linux.
Linux deployments, at least in the sector I worked with, were mainly Unix replacements.
Oh, and a couple responses to the above:
Advice: on VPS providers
Microsoft, ironically, tends get these sorts of wins as well. After all, everyone has Microsoft software sitting around. It's almost as easy to get rolling on a skunkworks Microsoft project as it is to roll one out with Free Software.
Well done dodging the vendor meltdown bullet, however. In my experience that basically never works. After all, it is pretty rare that a vendor can't point to other customers with successful implementations. Generally speaking when a customer has to flush a large investment down the tubes the guys that chose the tools and then were unable to implement the solution get run as well.
Let's just say I'm not a firm believer in the "throat to choke" theory of choosing software.
My real question for you is why did you move away from the less-expensive, integrated Microsoft solution that worked to something more expensive and less integrated. Nothing personal, but that doesn't sound like the sort of thing that any of the people I've ever worked for would blame on a vendor.