09.08.08
Stock Market.NET: Fail
From airline crashes to stock market crashes
Last week we used the Federal Aviation Administration (FAA) as an example of the unsuitability of Microsoft software for mission-critical operations. Today we witness another disastrous case study. It takes place in London where the stock market has crashed. [Hat tip benJIman for the pointer]
Trading on the London Stock Exchange has been halted after a computer system failed on one of the most frantic days of trading so far this year.
FTSE 100: trading on the London Stock Exchange halted after computer systems failed
In an embarrassment for the LSE, the exchange said that no orders can be entered or executions of those trades occur. The LSE plans to bring back trading in a “controlled way’’, but couldn’t say how long that will take.
Who is to blame? “Everything Microsoft” CIOs like Richard Steel immediately spring to mind [1, 2, 3, 4, 5] because they are very prevalent in the UK.
This London Stock Exchange (LSE) deployment is Microsoft’s wannabe ‘poster child’[1] , for which Microsoft claims “One hundred per cent reliable on high-volume trading days.” Can this ambitious statement be true given those previous outages [2]? LSE is also rumored to have excessive redundancy/backup, which raises costs considerably.
There are a variety of ‘peripheral’ ways in ways in which Microsoft software harms trading. For example, those hundreds of millions of Windows zombies enable stocks to be played [3]. This has already affected both the European [4] and the American market [5].
Reliability is not a strong point for Windows and other layers of the Microsoft stack. The Diebold fiasco is proof of this [6,7]. Meanwhile, wiser officers in the US market [8] and other international markets [9] are choosing GNU/Linux. It’s very pervasive [10, 11, 13] and even Free software has found its way in [12, 14]. GNU/Linux is said to be ready for mission-critical applications [15, 16, 17] and other bits of Free software are getting there too [18]. █
___
[1] London Stock Exchange Cuts Information Dissemination Time from 30 to 2 Milliseconds [DOC] [HTML]
Working with Microsoft and Accenture, the London Stock Exchange replaced its London Market Information Link (LMIL®) system with Infolect®. It used the Microsoft® .NET Framework, the Microsoft Visual C#® .NET development tool, and the Microsoft SQL Server™ 2000 database
[2] London Stock Exchange blames outage on Infolect
The outage hit shortly before close of trading and forced the exchange to extend its closing auction by a hour and a half, from 4.30pm until 6pm. Even so it was reported that that some traders left for the day with trades unexecuted
Infolect was launched two years ago in place of the exchange’s London Market Information Link platform. It uses Microsoft .net technology and a SQL Server database
[3] Microsoft is the world’s biggest cause of zombie remailers
In China, it would take about one and one-half years wages (for the average Chinese) to buy a legitimate copy of Windows Vista. If you could find it here.
Microsoft is the biggest cause of zombie remailers in the world, because they make noises, but do not do anything to address the real digital inequities in the world.
[4] Pump and dump spam comes to Europe
Earlier this month, the US Securities and Exchange Commission (SEC) had suspended trading in 35 companies that were found to be commonly referenced in pump-and-dump stock email campaigns.
[5] SEC goes after stock spammers, hackers
The SEC action to freeze the assets is the third filed in as many months involving market manipulation schemes conducted through online account intrusions. In the past the SEC has also taken action against individuals who allegedly broke into financial news Web sites or news release services.
[...]
“When spam clogs our mailboxes, it’s annoying. When it rips off investors, it’s illegal and destructive,” SEC Chairman Christopher Cox said in a statement.
[6] Ohio Audit Says Diebold Vote Database May Have Been Corrupted
Problems found in an audit of Diebold tabulation records from an Ohio November 2006 election raise questions about whether the database got corrupted during the tabulation of election results…
[...]
The database is built from Microsoft’s Jet database engine. The engine, according to Microsoft, is vulnerable to corruption when a lot of concurrent activity is happening with the database, such as what occurs on an election night…
[...]
The report mentions that election staff had trouble with the server crashing and freezing on election night.
[...]
The report notes that with punch card machines election officials used to be able to determine definitively if all ballots had been counted in the results.
[7] Diebold’s Problems Worse Than Reported, Tests Find
Nineteen machines had 21 screen freezes or system crashes, producing a blue screen and messages about an “illegal operation” or a “fatal exception error.” A Diebold technician had to restart the machine for voting to resume. Ten machines had a total of 11 printer jams. Almost one-third of all machines in the mock election had a problem.
[...]
“Especially with this blue-screen problem, you don’t know whether it’s the printer drivers, you don’t know whether it’s Diebold’s own code or whether it’s Windows, or where the problem is,” he said. “It brings into question the entire system.”
[8] NYSE Euronext banks on Red Hat
If anyone out there persists in believing that Linux isn’t ready for serious prime time, NYSE Euronext’s dependence on Red Hat should finally lay that silly notion to rest. As announced, the New York Stock Exchange Euronext dumped its proprietary UNIX heritage (AIX, HP UX, Solaris) for the freedom, flexibility, and performance of Linux.
[9] Linux steps up to power NZ Stock Exchange
The New Zealand Stock Exchange is moving to a Linux platform for its settlement and clearing system, replacing its existing HP NonStop platform and applications in order to reduce cost and increase flexibility.
[10] HP Linux servers bolster NYSE trading app
“We favor Linux for what we do. We don’t want to be beholden to any one [hardware or software] supplier, even if it is very good. We want the freedom to be vendor-independent, so Linux was a good choice,” said Rubinow.
Everyone is already using Linux somewhere, Elliott said, but there are still people who are afraid to put their mission-critical applications on it for fear that it is not robust enough.
[12] Open Source Slowly Gains Momentum on Wall Street
Customers may be comfortable using Linux for infrastructure-type systems but less so for business software, said Monica Kumar, senior director of product manager for Linux and open-source software at Oracle.
[13] Linux goes to Wall Street, puts on a show
“The Financial Services firms are leading the enterprise adoption of open source technology, including Linux,” says Zachary. “By watching the consumption patterns of these firms, open source vendors can gain a good understanding of the types of products and services that are commercially viable.
[...]
“We see a growing role for open source software beyond Linux in financial services companies and that’s why many companies are expanding the governance policies and procedures around using open source software.”
[14] Open source and mission critical: The Linux application server landscape
In 2005, Michael Dortch, executive editor and director of IT infrastructure management strategies at the Robert Frances Group, penned a report comparing Linux application server total cost of ownership (TCO) with Microsoft’s Windows and Sun Microsystems’ Solaris. In that report, Linux won handily, prompting Dortch to recommend that all IT managers at least give the Linux application server a serious look for mission-critical data center tasks.
[...]
What a lot of our clients are finding today, however, is that, with Linux and open source application servers, the wisdom of the crowd has solved those standards-based problems. And because the management tools are freely available and malleable, they find that the big sell was “cheap and free” but is now “What will complete my processes faster?” Open source used to be inexpensive and free, but now IT is finding value [in it] because it’s more responsive and responsive.
[15] Linux Crossing Threshold to Real Mission Critical Computing
…Linux developers are steadily chipping away at the most significant functional gaps that limited its suitability for these workloads in the past, allowing it to progress far beyond the “edge of network” domain where it has traditionally thrived. The impressive wins by Novell and Red Hat prove that user perceptions about Linux’s limitations are gradually crumbling as well.
[16] Next Gen, Mission-Critical Apps To Be Deployed on Linux Says Report
By now it should be obvious to even the most casual industry observers that Linux operating systems – and open source-based software in general – have reached critical marketplace mass. Recent Linux deals and announcements by Oracle and Microsoft have only reinforced the “open source is enterprise-grade” message that IBM, Unisys and other “Master Brand” hardware, software and services vendors have been preaching for years. In short, open source, especially Linux, is becoming “legitimized” by the major vendors for enterprise environments, and user executives are more than happy to believe them.
[17] CIO study finds Linux ready for prime-time
The company predicts a steep rise: only 18 percent of businesses will be using Linux in business-critical roles by the end of 2007.
“Linux operating systems – and open source-based software in general – have reached critical marketplace mass,” said the study’s authors, Bruce Guptill and Bill McNee of Saugatuck Research.
[18] Forrester: Open source is everywhere, and increasingly used for mission-critical applications
Forrester just issued an insightful report on the increasing adoption of open source in the enterprise. The verdict? Open source is everywhere, and taking an ever-increasing piece of the enterprise pie
Needs sunlight said,
September 8, 2008 at 1:46 pm
Two more Windows incidents:
http://www.techworld.com/opsys/news/index.cfm?NewsID=2275
http://www.engadget.com/2007/08/16/network-card-crash-leaves-17-000-stranded-at-lax/
The license (if valid) disclaims both liability and suitability, then the fault for the above can be placed *solely* at the feet of those responsible for letting M$ getting even a toehold in at work.
Roy Schestowitz said,
September 8, 2008 at 2:04 pm
Thanks. I collect these and will occasionally post a roundup.
Jose_X said,
September 8, 2008 at 3:09 pm
>> The license (if valid) disclaims both liability and suitability, then the fault for the above can be placed *solely* at the feet of those responsible for letting M$ getting even a toehold in at work.
Excellent point. Assuming FOSS could find no backer to offer any further guarantees than as stated for Microsoft, I think it would make lots of business sense to really want to have access to the source code. Only you care about your business as much as anyone could. Microsoft, with their many many other large clients certainly isn’t going to call people out of vacation for you, especially if they have little to lose: “there is always the client that mishandled our software”, “they can’t successfully sue”, “they are locked-in and will come back for more”, “they represent a small portion of our business”.
In this particular case, Microsoft will likely work hard to help solve the problem, but for the typical user out there that goes with Microsoft products, expect less than little real help from Microsoft compared to what you could get with source code access and free or paid online sources and a smaller more dedicated service provider.
The bigger problem with going with Microsoft is that you don’t know what kind of business intelligence Microsoft is grabbing from your business. Innocent stats can go very far in giving Microsoft and any partners willing to pay more than you an advantage over you.
I also see so-called malware writers gaining advantages from the holes in Microsoft products that you don’t and can’t know about or fix in a timely manner. With Linux, you can see and fix relevant issues that crop up even before complete failure. You (or those who really want your business, in competition with each other) can manage risks and resources when you have access to the source code of all the software running your business. It’s smart to know.
Michael said,
September 8, 2008 at 7:22 pm
Well if this doesn’t get someone ‘fired for choosing microsoft’, i don’t know what will. And so much for the total BS of ‘one throat to choke’ all the lecherous middle management like to spout.
Is the LSE going to sue any of their suppliers or just take the fallout themselves? Mission critical stuff on .NET? Then again, they deserve whatever they end up with.
Roy Schestowitz said,
September 8, 2008 at 11:08 pm
They blame connectivity, but that too is part of the overall system.
Roy Schestowitz said,
September 8, 2008 at 11:52 pm
I found this:
“Traders said the exchange was suffering an ‘overloading’ caused partly because the FTSE 100 ended down sharply on Friday so traders who were betting on further falls in the index would have to close their positions.”
So connectivity was just a poor excuse if you ask me.
Needs Sunlight said,
September 9, 2008 at 1:45 am
@Jose_X:
I have *heard*, but not confirmed, that MS licenses aim to prevent a lot of legal choices. Some of the more bizarre clauses suggest that the customer must back/support MS in court.
If you can manage to get a printed copy of the *all* the licensing agreements relevant to such a heinous and convoluted stack as MS supplies in situations like for LSE, then you might find all kinds of material to prevent acquisition of MS’ imitations of technology.
That is if the licenses and contracts are not under NDA. If that is the case, then that too needs to be brought out into public scrutiny.
Needs Sunlight said,
September 9, 2008 at 1:47 am
@Roy:
No. They do not blame connectivity, they blame the ‘network’, but without indicating the context in which the word is used.
Same words and tactics as used in whitewashing the Sampo Pankki catastrophe.
Sure it’s mincing words, but do not read into it something which is not said.
Roy Schestowitz said,
September 9, 2008 at 2:04 am
I read an article where some reps blindly tries to defend the system’s integrity. I believe that they will never come out with a real report (with or without a consensual ‘gagging order’).
Jose_X said,
September 9, 2008 at 7:15 am
>> This London Stock Exchange (LSE) deployment is Microsoft’s wannabe ‘poster child’[1] , for which Microsoft claims “One hundred per cent reliable on high-volume trading days.” Can this ambitious statement be true given those previous outages [2]? LSE is also rumored to have excessive redundancy/backup, which raises costs considerably.
And…
What is the cost of missed business? What is the cost of a damaged reputation? Will Microsoft’s “independent” client update the TCO studies for the LSE? Do they know what TCO is?
Is Microsoft advertizing false material?
Not sure if these contain any new info, but I’ll post the links anyway:
http://www.bloomberg.com/apps/news?pid=20601087&sid=a9bRidj83NBQ&refer=home
http://www.reuters.com/article/ousiv/idUSL01084620080908
Roy Schestowitz said,
September 9, 2008 at 7:33 am
Does the sun rise in the east?
Seriously now, this is not the first such incident and this never happens in other markets that run ‘legacy’ systems, including Solaris and GNU/Linux.
Netcraft has just crowned GNU/Linux for uptime (again). More on that later.