London Stock Exchange To Abandon Windows 438
BBCWatcher writes "Computerworld's Steven J. Vaughan-Nichols reports that the London Stock Exchange is abandoning its Microsoft Windows-based trading platform: 'Anyone who was ever fool enough to believe that Microsoft software was good enough to be used for a mission-critical operation had their face slapped this September when the LSE's Windows-based TradElect system brought the market to a standstill for almost an entire day .... Sources at the LSE tell me to this day that the problem was with TradElect ...'"
Elementary, my near noob (Score:2, Insightful)
Anyone who was ever fool enough to believe that Microsoft software was good enough to be used for a mission-critical operation had their face slapped this September when the LSE (London Stock Exchange)'s Windows-based TradElect system brought the market to a standstill for almost an entire day. While the LSE denied that the collapse was TradElect's fault, they also refused to explain what the problem really wa.
Right, so it wasn't M$'s TradElect's fault, therefore it clearly was M$'s TradElect's fault. Someone give this guy a job at the FBI!
Not Windows' fault (Score:5, Insightful)
It's not Windows vs Linux.
It's TradElect vs MarketPrizm, which happen to run on Windows vs Linux respectively.
TradElect never managed its performance promises, which suggests lies from marketing and / or programmers unable to deliver what they were asked to. Despite what the Linux fanboys love to say, inferior software isn't Windows-only, and does exist on Linux too.
This could easily have been the other way around, ditching Linux and a shit piece of trades software for Windows and a good bit of trades software. The OS is irrelevant here, except to fanboys of either side.
Re:Seems like a bunch of unknowing (Score:3, Insightful)
Btw. the same would apply to a blank linux java stack....
You need realtime stuff to do that!
Re:Not Windows' fault (Score:5, Insightful)
Exactly. We run 90% of our mission critical software at work on Windows and don't have problems...but have moved to Linux to run remote monitoring software. Should I make a post saying we're abandoning the horrible crappy Windows market for Linux? That would be just as incorrect of a statement.
Come on. TradElect is the problem. (Score:5, Insightful)
Re:Not Windows' fault (Score:5, Insightful)
It's not Windows vs Linux.
You say it's not Windows' fault and I agree--it wasn't an OS problem (per se), but rather an application issue. In actuality, it's Microsoft's fault; the application was developed in joint by Accenture AND Microsoft. With the requirements not being met that it be a high-performance, real-time application and the fact that they were unable to deliver even with MS being involved made them lose faith in the company and their products (.NET, Windows Server, SQL server).
I'd say that if MS wasn't involved in the development of the app that it's possible that they would scrap the app rather than the OS/framework, but if I was in that position, I'd do the same thing.
It's possible that they also look at the chicago stock exchange and the NYSE and the fact that their apps are running on Linux and have decided to move to a proven, successful system.
Re:Not Windows' fault (Score:5, Insightful)
Then again, TradElect was written by Microsoft and Accenture [onwindows.com], so Microsoft where heavily involved in this project themselves - not just from the perspective of Windows only.
In addition, they touted this in their "Get The Facts" anti-Linux campaign, so I'm sorry, but pointing out this failure and blaming it on Microsoft (though perhaps not the Windows OS as such) is fair game IMO.
I mean, if a large and well-known consulting firm together with Microsoft themselves can't make a Windows-based framework perform, who can?
Re:Not Windows' fault (Score:5, Insightful)
No, it's Microsoft vs Linux.
Microsoft had full control over the stack of tools they used (Windows Server 2003, C#/.NET, Sql Server 2000, I believe) and they invested a lot of resources, both technical and marketing, into making this system run. It was suppose to show that Microsoft software could handle this kind of system as well or better then *nix. And it was a failure.
See Get the Facts [microsoft.com] for more details.
Oh come on. The article is a troll. (Score:2, Insightful)
IF the article had said "We looked into this system in some detail, albeit after we'd bought it, and discovered that anything based on Windows was fundamentally incapable of meeting our needs. So we've decided to move", then I would take it more seriously.
As it is, it sounds like a political move from a new managing director who's trying to make themselves out to be as different as possible from the previous one - and one of the ways they're doing that is to ditch the computer system. The fact that the old system runs Windows and the obvious alternative runs Linux is neither here nor there.
Face facts, it's just as possible to produce a lousy system based on Linux as it is on Windows.
Re:Come on. TradElect is the problem. (Score:5, Insightful)
The fact that not even Microsoft’s involvement was able to make the system meet its requirements looks *very* indicative to me.
Maybe not Windows fault but sure is Microsoft's (Score:3, Insightful)
From Microsoft's case study http://www.microsoft.com/casestudies/Case_Study_Detail.aspx?CaseStudyID=200042 [microsoft.com]
In the development, roll-out, and implementation processes, Microsoft worked closely with the London Stock Exchange to ensure not only that they understood their immediate requirements, but that the solution fitted their long-term business plans as specified in the TRM project.
Microsoft was equally involved in this project no matter how you try to spin it.
Re:Let me be the first to say... (Score:4, Insightful)
(Of course I know I wasn't... but cmon!
Re:Not Windows' fault (Score:5, Insightful)
The OS is irrelevant - every modern server OS performs well enough to support sanely written software and sanely designed infrastructure. Only the people living in the past and the ones having no clue will argue otherwise.
Or the ones who don't know what they're talking about, like you.
This is one of the main stock exchanges in the world. Billions of dollars of trade rely on microsecond-precise handling. There are whole companies (and not small ones) that do stuff like inter-exchange trading which is the buzzword for "buy for $1,5678 in London, sell for $1,5679 in Tokyo before anyone else does and the prices equalize". These are companies that are willing to put down five to six digit sums per month if they can get an Internet connection with a few milliseconds less latency.
For this environment, you don't need "sanely managed". Any delay whatsoever in the transactions is bad. Any time a transaction can not be handled properly due to delay, queues or any fucking other reason, one of your traders is unhappy. And you don't want unhappy traders when they are your business.
Re:Elementary, my near noob (Score:5, Insightful)
I'm a bit confused as to why it's Microsoft getting knocked on the head here. Sure, SQL Server 2000 might not be the best choice, but how are we to know what actually caused the issues? You could write poorly written code anywhere, and outages could well be caused by hardware failure and poor failover planning. To blame it all on the .NET framework seems a bit odd to me, without knowing what was causing all the problems.
Of course, I'm not a big fan of SQL Server databases for huge mission-critical applications (multi-version consistency in TempDB version stores, anyone?).
Re:Not Windows' fault (Score:5, Insightful)
I've always believed that Microsoft don't really get mission critical software so I'm surprised they got the contract. My experience with their OSs suggests that time and time again they fail to get the basics right or that things just work superficially. They cover this up by submerging it in a slopping sea of unwanted bloaty features.
What this implies is that they must have damned good sales executives to overcome the word on the street.
Re:Not Windows' fault (Score:3, Insightful)
They have to reap what they sow here. If you use someone else's work as an example of your own ability, you better be damned sure you understand it's quality.
Re:Not Windows' fault (Score:2, Insightful)
You say it's not Windows' fault and I agree--it wasn't an OS problem (per se), but rather an application issue. In actuality, it's Microsoft's fault; the application was developed in joint by Accenture AND Microsoft.
Being currently an Accenture employee, i find it increidble that the application managed to work at all. Sometimes i wonder if they get tax cuts from hiring incompetent people.
Sorry i have to post within the shroud of anonymity.
Re:Let me be the first to say... (Score:5, Insightful)
But, then, it's not often you see enterprise software fail quite so badly and publicly as was the case with the LSE
A quote from another source is appropriate here:
This is a good death. There's no shame in this, in a man's death. A man who has done fine works. We're making a better world. All of them - better worlds.
article:
So, might I suggest to the LSE that they consider Linux as the foundation for their next stock software infrastructure?
Re:Not Windows' fault (Score:5, Insightful)
I seriously doubt microsoft was involved in the development of tradelect. marketing in collaboration with accenture yes.
From an old Computer Weekly [computerweekly.com] article
"Accenture built the Tradelect platform in India between late 2004 and March this year."
And from an old information age article [information-age.com], a classic Quote from the now departed IT Director:
"That was where Microsoft came in. We looked at their whole suite of technology from their development environment through to their databases and operating systems, and we decided that their technology was best aligned to achieving this range of design principles. We also found that they were willing to operate as true partners with us and to engage throughout the whole four-year programme rather than on particular components within it where there was potential revenue for them through licence sales. So we felt that not only did their technology stack up against the design principles, but they were genuinely able to act as a partner. They recognised at the most senior levels what we were trying to achieve here and that was important to us."
That's £40m over a short 2 years of service [onwindows.com] - work out the TCO on the depreciation cost alone! So, yes, I do think Microsoft has a lot to answer for because they were engaged at the highest levels. Also, Accenture have a lot to answer for. As soon as I saw "India", well, I'm sorry, but it's rare for an offshore project to meet requirements - in the same way that a project for Bank of India outsourced to the UK would probably fail.
It's worth a look at the Chi-X platform sales brochure [sii.org.uk] (it's PPT, how ironic) which is a direct competitor to LSE and uses Linux successfully. Chi-X has about 15% or so of UK FTSE 100 trades. The amazing feature of CHi-X is its low latency - especially in trading where 20 ms is a very long time and can cost principals serious money.
are you fucking blind ? (Score:5, Insightful)
Mulitple Problems (Score:5, Insightful)
I was involved in discussions (and more) with the LSE before (during but not so much) and after they decided to select Microsoft / Accenture / India / Outsourcing as the path for their solution and I know some of the key decision makers. Under the Microsoft umbrella, they were significantly influenced by the resources Microsoft was willing to commit to making the project work despite the newness of .NET as an ifrastructure.
It is important to remember where the LSE was before the TradeElect project, they had completely outsourced their platform to Accenture, the amount they spent per annum on keeping that platform up and running were phenomenal, an order of magnitude more than some of our clients were spending and they (our clients) were running much higher performance systems. TradeElect was designed to decrease these costs without compromising the "I don't lose sleep at night worryin about the systems" position of senior managers. I firmly believed it was a mistake to believe that .NET at the heart of the platform would meet the requirements of an exchange trading platform.
I have no real issue with Windows as the OS under the platform, really for a trading system the OS is providing a TCP stack and some IPC and thats about it. Everything else and the vast majority of the bottlenecks are in your application stack, whether it be tools or application code you are writing for your specific problem domain. Although one might argue that the Microsoft IPC tools can be argued as "weak/complicated".
It will be interesting to see which people the LSE use to provide the analysis of which way to jump with this decision. Too many very senior folk were involved all the way through the TradeElect project for heads to roll, but it will be interesting to watch who says what when the final decision of what to do is announced.
Re:Not Windows' fault (Score:4, Insightful)
The point you are missing is that the application is not independent of the operating system. If the operating system does not support, or supports poorly, such features as real time processes, >1 CPU core threading with appropriate parallel locks on IO buffers and channels, O(n) process and thread scheduling, redundant failover, and a thousand other small features, then implementing applications that are high performance, real time, and mission critical, is massively more complex, or even impossible.
What actually runs on the CPU is application+OS user space libraries+kernel+kernel drivers. The performance of the sum total is as dependent on the performance and capabilities of other every bit of the stack as it is on the application itself.
Re:Crash over Switzerland (Score:1, Insightful)
Switzerland scenario + idiot pilot = crash
Switzerland scenario + smart pilot = no crash
Airbus scenario + idiot pilot = crash
Airbus scenario + smart pilot = crash
Car analogy! We may get upset at auto companies that take shortcuts in safety features that may cost lives when the cars crash (Pinto), but I have never seen anyone blame Ford for the accident occurring in the first place.
Out of curiosity, who made the planes that crashed over Switzerland? You seem to imply Boeing but don't explicitly say it.
Re:Your source is from 2005 - Mine are from 2006/2 (Score:2, Insightful)
You think MDDS is their trading system? It is not
Supermontage is what is executing NASDAQ trades (http://www.investopedia.com/terms/s/supermontage.asp)
The citations I posted were older because that was when Supermontage was rolled out.
Your FUD rolls off me like water on lotus leaves
Here is what actually does the work you claim MDDS does (http://www.tibco.com/resources/customers/successstory_nasdaq.pdf )
Re:Not Windows' fault (Score:2, Insightful)
Did you ever stop to think that there might be something fundamentally wrong with a financial system where fortunes are made and lost based on a split-millisecond of "trading time"?
After all, the actual value is entirely created by some farmer raising hogs in Iowa and someone else mining nickel in Sudbury.
Re:Not Windows' fault (Score:3, Insightful)
It's always Microsoft's fault.
Not always, but when Microsoft drives your trading platform and it crashes this badly, it sure as hell is.
Re:Not Windows' fault (Score:2, Insightful)
It's really good that it works well for you!
Really!
However, as the masses can testify, that is not the norm. The complexity in the O/S makes it very susceptible and prone to conflicts. Some people argue that you need to know how to make it work properly, which is true, but should not require that amount of dedication just to have some service up and running.
In the 90's I had a 486 w 8 MB RAM run as my web server, DNS, mail server, SQL server and acted as a router with a firewall on a dialup connection. Later I added another 8MB so that I could also run the GUI.
It's due to the superior simplicity of the UNIX design.
Windows is heavy on complexity, thus less manageable, and more prone to problems. Clearly a lot of people have a working Windows setup, but I'd say many more people have massive stability issues. UNIX was not designed with security in mind, but because of the simple modular design can be modified to work with it in mind.
Now if you don't work with different O/S over time in environments where you can really compare O/S's handling different situations you might not realize how much easier it is to, for example, manage Linux servers compared to Windows. I think the number is something like 12 servers per admin. You could possibly reach 100 servers per admin under a UNIX like O/S. Again, due to simplicity over complexity.
For those of us who have had the pleasure of knowing other O/S's, we can testify to the pleasure level of getting off Windows. For those who love Windows - good for you! It has certainly gotten better over the decades. Is more stable and so on.
Re:Oh come on. The article is a troll. (Score:3, Insightful)
The reason this is significant is that Microsoft were heavily involved in the development of this system and are still saying how wonderful it is that the LSE runs on their amazing software. The fact of the matter is that it's been a total disaster both for the LSE and for Microsoft's PR machine. Whether it's a political powerplay as well remains to be seen, but the simple fact is that the software sucked and not even Microsoft could make it work.
Re:Let me be the first to say... (Score:3, Insightful)
I have not met a customer in the last 20yrs who did not use windows as the OS for all or part of their system(s).
Re:Crash over Switzerland (Score:5, Insightful)
Re:Similar at several European banks (Score:3, Insightful)
Which dosn't help that much if the actual problem is that Windows isn't the right tool for the job. It would be a bit like someone in Florida congratulating themselves on being able to buy snowcats at low prices
Re:Not Windows' fault (Score:2, Insightful)
Etc, etc... Yes, it can all be done with open source/linux software, but to get it all to work, well - you're looking at HEAPS of custom in-house software or product choices to make.
I've been using Windows since 1989 and Linux/other unix since 1994, and yes both are getting better. However until linux addresses things like the above, it will remain confined to specific niche deployments or power users. Most of that goes for OS/X as well.
Which OS runs 5-10% faster or uses 20% less ram is no longer really relevant. Its how well it can be centrally controlled and managed across large numbers of machines. Linux is not there yet; I do wish people would start focusing on that though, but its going to have to be fucking impressive to get business users off active directory.
Re:Exchange Server (Score:3, Insightful)
Tell him to have less meetings, so people won't have to compete for rooms, projectors and so on.
Productivity will soar.
Give programmers private offices with a part of the profits.
Re:Not Windows' fault (Score:1, Insightful)
In addition, it's almost certain that he not only has no current working knowledge of Windows servers and AD (I still think that eDirectory is more robust, and scales better, but AD certainly has come a long way - we're testing Server 2008 R2, and so far I'm impressed for the most part), but that he doesn't have any awareness that the things you mentioned in your post are issues at all, because for him they are not. He's probably gone from that 486 to a modern PC, running all the same stuff, pretty much, and thinks that what works for him automatically applies to any network in the world, regardless of size. It'd be funny, if it weren't for the fact that there are so many like him on Slashdot now, narrow-minded OS bigots that see the world in absolutes, and need to think that their chosen OS is best in order to bolster their egos.