London Stock Exchange to abandon failed Windows platform

The truth is like the sun

You cannot hide it with your hand.

Hardware and software license cost insignificant

When are Linux proponents going to realise that harping on about free licenses and cheaper hardware is pointless? In the average enterprise IT project, these costs form only a tiny proportion of the deployment and support costs.

When you factor in the use of RHEL (most enterprise's chosen Linux platform), then the software support license costs don't really bring a financial advantage over Windows servers.

The only thing that matters is the business capability of the deployed systems - it makes absolutely zero difference to the business whether they run Red Hat, Solaris or Windows 98, as long as the capability is there and reliable at the lowest cost. If its cheaper and just as reliable to push out a pool of 20 Win2k3 servers than a pair of Solaris boxes, then why choose Solaris or *nix?

I think you answered your own question...

"If its cheaper and just as reliable to push out a pool of 20 Win2k3 servers than a pair of Solaris boxes, then why choose Solaris or *nix?"

Well, because it isn't cheaper, and it SURE as heck isn't as reliable.

When you go Windows, not only do you have your server and desktop licenses, but you have your client access licenses for every device connecting to the server itself, for Exchange, and for SQL server, etc... and then you've got to have licenses for your antivirus on all those machines, and software assurance if you don't want to get screwed on the upgrade, and you have to have more machines to do the same work, and these machines have to be more powerful because Windows is a resource hog, so you spend more on electricity and cooling.... jeez, the costs just keep adding up! And then you call for support, and you get charged a ridiculous rate. Oh, and then you need to pay $300 a machine for MS Office licenses. So there goes your cost argument...

And I don't think I even have to start on the reliability argument -- if you think Windows is even AS reliable as a *nix, you're a step away from a home for the mentally disturbed.

Not exactly

The Exchange is run on many 2003 servers, but there is most definitely a volume license for both the servers and the SQL instances. Also since a mission-critical environment is needed there is more than likely no antivirus software required. These servers I'm sure were very finely tuned with respect to resources, etc. It's not very difficult to get a very fast stable/reliable 2003/2008 server running very few services – turning on only what is needed.

_ _ _

And I don't think I even have to start on the reliability argument -- if you think Windows is even AS reliable as a *nix, you're a step away from a home for the mentally disturbed.
_ _ _

And if you really think that, you are incredibly ignorant.

Sticking with a costly decision

It is difficult to abandon a software system after it is implemented. It must be a 'massive fail' in order to acknowledge that is it crap and remove it. I've been in situations where the software chosen has been utter crap, yet we are expected to make it work. After months or years of problems, only then will the decision makers re-visit the question, to keep or purge.
GNU/Linux is the path to productivity.

LSE management was stupid

From TFA:
"...the CEO that brought TradElect to the LSE, Clara Furse, has left without saying why she was leaving..."

So they are getting a platform change and a management change. From my experience there are several Microsoft shops that will require this type of intervention before they can break the amazingly expensive cost of running Microsoft sub par software products.

CEO v.s. CTO

Thats right the CEO left not the CTO. The CEO made a bad spending decision with regards to the technology they used.
The CTO was the one that probably proposed to use the M$ crap in the first place.
This still means that LSE could still F&*% up even with Linux as their platform.

This was a management call

The LSE outsourced the whole thing from soup to nuts to MS and Accenture.

LSE's in-house IT was Not amused.

I strongly suspect that this was one of those sweetheart deals where the customer pays less and MS got to trumpet how the project showed how enterprise-ready the MS software stack was.

Except, of course, as events have shown, it was anything but enterprise ready.

Frankly, the project appears to have been mismanaged from moment one. While I certainly would never have built anything like this on top of MS software, they could have made it faster and more stable just using other MS programs. I mean SQL 2000!? SQL Server 2003 was already out and battle-tested by the time this project was under way.

While I like to kick MS around when they blow it, the integrators on this one probably should take the lion's share of the blame.

Steven