London Stock Exchange hauled offline after major data problem

London Stock Exchange hauled offline after major data problem

Exchange prices continues to be subject of technical issues

Article comments

The London Stock Exchange stopped trading this morning immediately after the opening auction, which experienced a major technical glitch.

The news comes at the end of the second week of trading on the exchange’s new Millennium trading system. Over the last fortnight, anger has been growing after large vendors that supply price data to the market continually experienced significant technical difficulties, including displaying entirely inaccurate and blank prices.

The LSE said the nature of the technical problem had not been established, though it told traders of a “market data issue”.

Best Bid Offers have been the subject of numerous difficulties throughout the last fortnight. Several vendors have highlighted to Computerworld UK that cancellations of orders had presented them with data effectively “stuck” in the system and blocking the correct transaction of other orders.

The problem was first announced at 7.54 this morning, when trades apparently cancelled on the system.

“The uncrossing trades that took place in the opening auction today are under investigation for potential cancellation,” the LSE said.

As of 8.59, the LSE said uncrossing and automatic trades from securities this morning will stand. Orders can be deleted, but not executed manually or automatically.

Millennium Exchange, the platform on the main venue of the LSE, is written in C++ language and running on Novell SUSE Linux-based datacentres. It offers data through the FIX and ITCH protocols.

Yesterday, the LSE experienced pricing problems on a separate platform, SOLA, for its derivatives exchange EDX. The day before, its Milan exchange, Borsa Italiana, was offline all morning following major issues on TradElect, the .Net-based system dumped in London.

Now read:

LSE calls crisis meeting with market data vendors

After four hours, exchange is back online with excuses but no real explanation

Share:

Comments

  • josex And it is true that it might have been the hardware or simply an honest failure due to complexity somewhere however Linux does have a reputation for being faster than windows and holding up better under server stress eg 90 of 500 fastest supercomputers run Linux by choice We also cant deny that as long as the software source code is proprietary the number of possible interactions between OS and app or between any two pieces of software can be very large in number and be very difficult to discover
  • josex gtgt Josex might be wrong or heshe might be right but at least its a plausible hypothesisYes there could be many reasons but one can easily underestimate the problems when you change part of a software stack distinct components of software that up to that point were working together as one unitFor how many years has Microsofts Internet Explorer not faithfully adhered to web standards Ask those who build complex websites and sometimes even simple ones if it matters which browser is used to access the dataSo are we to believe that this particular client-network application is so different especially right after a changeThere might be various reasons but pretending we can overlook this issue of interoperability across different manufacturers especially when one of them is Microsoft is very possibly to fail to take this problem seriously or to be in denial
  • josex As a software developer I think it is a lot more than merely plausible but at least raising this subject in a public forum might help raise a little bit more awareness among those with less experienceThe LSE very likely has already considered this but the mere fact Microsoft has so many clients means that many software users likely dont realize how tricky it is to get interoperability correct across complex software stacks especially when the participating firms are actively developingupdating their software and more so when there is so much money on the lineOpen source is the only place where you find complex software created by many different folks able to run smoothly together In fact there are many open source stacks that run smoother together than even some Microsoft products running with other Micrsoft products The reason open source created independently can potentially interoperate so well is that in truth none of these folks are independent from each other because the source code is open to everyone to inspect as bugs come upTake a look at that ODF link from robweircom that I posted earlier for disclosure Rob works for IBM This is one example and where many people were watching Its also an example where the protocols were not created by Microsoft and were very open The reader might even want to research Microsoft and embrace extend and extinguishNote how Microsoft took an active role in developing the LSE trading software Whatever the reason you would think there are plenty of experts out there on Microsoft software and who work for different firms or are there rhetorical question If it takes experts from within Microsoft to figure out Microsoft software and protocols lets assume then dont be surprised if when you change part of the stack you find problemsI would suggest that Microsoft be brought in to help sort this mess out but I have already laughed pretty hard this week and still need to recover a little moreIt naturally works to Microsofts advantage every time someone tries to replace Microsoft software and problems ensue This is why it really only works to avoid Microsoft from the beginning or else prepare yourself for something of a battle later on
  • josex gtgt do you really think people connect windows xp to the core of a stock exchangeI dont knowUmmmI thought THE core stock exchange for the LSE had been windows-based until nowIf you want me to believe that the main exchange can run MS dotnet but the lesser units that hook up to it would not consider such a thought then you will have to argue that point a little bit or maybe a whole lot bettergtgt a collection of the top IT people in the world wouldnt have thought to try reinstalland remove their old data and application session state as well I dont know I can see how many people might not Most people keep data especially app data and personal data when they reinstall WindowsOf course I was giving Microsoft a bit of the benefit of the doubt as to their intentions Microsoft has and easily can create problems In fact interoperability of very complex software is not easy to achieve without extensive desire and effort This is one reason why lock-in has been such a popular component of the business model of software firms writing proprietary software Its also why open source software has attracted so many developers that actually want to interoperateIts normal for many bugs to be created by developers in the same firm and there will be many many thousands of them any of which can interact in complex ways with other parts of the system under various conditions however when the same firm builds multiple tools and interacting software many of those bugs carry over into related tools because there is code reuse There is also extensive testing among components produced by the same firmHere just look at one single example of how badly Microsoft interoperates even when the standard is very open ODF and there is pressure on them to do so httpwwwrobweircomblog20 How can a firm with so much money and supposedly talented developers as should have Microsoft be so sloppy when it comes to interoperability It could be that Microsoft has greater sloppiness than you see in competitors but Ill tell you this much Since the 80s Bill Gates and friends have been a bit bonkers about competitors to DOS and other core Microsoft software Microsofts monopolies the inability generally of third parties to replace MS software within a stack is the reason why so many billionaires exist today with roots in Microsoft So tell a billionaire to be less sloppy so that he can instead of being a billionaire be a millionaire and tell me what he says
  • Guest I believe the real problem is not Linux but that the LSE bought in an inexperienced and unknown Sri Lankan company called Millenium IT to provide their Linux trading system Millenium IT was a small Unix reseller which set up the miniscule Colombo stock trading system - its only experience of trading systems Millenium IT seem to have been very good at selling themselves to LSE but pretty abysmal at everything else It is interesting how their lack of competence seems to extend to the point where they cannot even trace what is going wrong with their systems First they blamed sabotage which seems to have no foundation then they hired a lot of C programmers - which appears to indicates bugs in they software as the cause and now they do not seem to have a clue about what is wrong with their systems in the latest failuresThe reason LSE had for buying in an inexperienced untested and unreliable supplier and system rather than buying in proven reliable and faster performing trading systems from suppliers which have supplied other bourses Linux based systems and have been running faultlessly 247 at higher capacities for about 10 years now is simple - LSE wanted a proprietary system they thought they could sell to others and fancied themselves as software developers Millenium IT was the only one they could buyMy advice to LSE is stick to your core competency which is finance dump the Millenium IT system which is no better and possibly than the failed Turquoise NET system and employ competent external IT suppliers like the ones who supply the excellent NYSE or Deutche Bourse systems
  • Some Try reading for comprehension Look up terms you dont understand eg client sideThink then writeJosex might be wrong or heshe might be right but at least its a plausible hypothesis
  • TestSkhan Dont fix it if it aint brokenif you want to then do so with time enough time to testI would blame the project manager
  • anonymouse do you really think people connect windows xp to the core of a stock exchange and a collection of the top IT people in the world wouldnt have thought to try reinstallin all likelihood all the client connections are linux server and have been for many years
  • TGM Please re-read your post youve compared an operating system Linux with an application the original Net system
  • Alan Gramont Wait Linux people are quick to blame every issue on Windows Now that Windwos is out of the picture suddenly its not the OS anymore Get real If it isnt Linuxs fault for this it probably wasnt the original Net system either Its probably just a complicated system that is trying to find a silver bullet for a network that just isnt able to keep up
  • josex Dont be surprised if the problem is with client side Windows machines using dotnet applications that dont speak the protocol properly Microsoft has a long long LONG reputation of doing protocols their own way and not by standards If the LSE was using dotnet before its very possible the protocol was done differently when the client and the server were both dotnet Additionally such clients dotnet software might have saved its history and other configurationsaved state information in such a way that now the problems are showing upOf course what I mentioned might not be the problem at all but it seems very plausible Reinstalling the client software possibly including Windows itself since that might store metadata used by the apps and discarding old data may solve the problemAt a minimum I would like to know what software stack is being used by those customers or customers of customers who are experiencing problems
  • Thameera I wont be surprised if everyone starts blaming linux for this
  • Was come on lads you are not playing board games hereAdmit your mistakes and ask for help
  • MM Well i think a 15 month testing window should have been plenty of time for the exchange get it right
Advertisement
Send to a friend

Email this article to a friend or colleague:


PLEASE NOTE: Your name is used only to let the recipient know who sent the story, and in case of transmission error. Both your name and the recipient's name and address will not be used for any other purpose.


We use cookies to provide you with a better experience. If you continue to use this site, we'll assume you're happy with this. Alternatively, click here to find out how to manage these cookies

hide cookie message

ComputerworldUK Knowledge Vault

ComputerworldUK
Share
x
Open
* *