[dundee] London Stock Exchange Drops Windows Systems

[Rick Moynihan]

2009/7/5 Iain Barnett <iainspeed at gmail.com>:
> 2009/7/5 Rick Moynihan <rick.moynihan at gmail.com>
>>
>> My first response was to this was, WTF they ran windows?!  And then I
>> remembered Microsoft touting this in their FUD spreading get-the-facts
>> campaign against Linux.
>>
>>
>> http://www.pcworld.in/india/news/7143439/Windows/London_Stock_Exchange_Drops_Windows_System
>>
>> Doubtless they'll soon be running Linux like every other big exchange.
>>
>> R.
>
> ...
> While the LSE denied that the collapse was TradElect's fault, they also
> refused to explain what the problem really wa. Sources at the LSE tell me to
> this day that the problem was with TradElect.
>  ...
>
> Surely, without knowing the cause of the downtime or his sources, this
> counts as Linux FUD (ooh, Windows will break, can't tell you why it broke
> here, but my mates who run the nix servers at the LSE blame Windows... )
> *yawn*

I largely agree though still find the case interesting.  Specifically
because Microsoft were so heavily involved in this very high profile
project, and were so keen to tout it as an example of the Window's
stack's superiority.

Obviously massive software projects such as this are prone to epic
amounts of FAIL, so it's likely as much a failure of project
management as anything else....  When projects like this, at their
inception set arbitrary requirements like being a pure MS/Windows
stack, you have PR and marketing making what should be technical
decisions.

I guess my point is, that large systems/applications like this have
huge scalability requirements and are essentially always an exercise
in specialisation...  There is no one-size fits all solution, and I
doubt very much the efficacy of any solution based on a single vendors
product line.

You might argue that FLOSS is in many ways just as bad, making
technical decisions on other arbitrary criteria (e.g. philosophy); and
this certainly occurs...  However, the OSS stack is flexible enough to
incorporate specialisation at every layer with technologies from a
massive diversity of vendors...   I'm betting the Linux based system
features a whole slew of specialised solutions, custom kernels and
likely some proprietary tech e.g. Oracle/DB2...  It's for these
reason's that I think it sounds like a much better alternative.  i.e.
the technical decisions were likely made on the basis of
technical-merit over marketing.

Obviously this is just speculation... Hopefully the real reasons will
someday emerge (but I doubt it).

R.