Murdoch hacks grumble over outsourced IT failures
Freddie Starr ate my email server
Posted in Management, 9th April 2010 11:15 GMT
Watch Now : Virtual Machine Movement with Hyper-V
To picturesque Wapping, where unrest among hacks and techies about recent IT cuts at News International is bubbling following a 24-hour email outage yesterday that crippled newsrooms and commercial operations alike.
Cast in the role of villain is CIO Andrew Hickey, the man responsible for a broad IT outsourcing deal in December that has seen many functions taken over by HCL, of Noida, near New Delhi.
The Reg's embedded correspondent sent this exclusive report: "The new disaster recovery system is a complete clusterf*ck.
"When the problem happened it replicated to all the active directory servers, so the email disaster recovery system was deleted completely."
Hickey has a background with Accenture, and has brought in several of his former colleagues to manage the soaraway outsourcing bonanza.
"He believes all IT departments from banks to supermarkets to newspapers can be run in the same way: ie just using off-the-shelf software administered by low-skill employees," our man says.
Editorial colleagues are also grumbling loudly about the decision to make deep cuts to the communications functions of a communications company.
Announcing the deal, Hickey - who clearly doesn't share Sun subs' gift for direct language - said: "I believe that longer term, technology cost optimization, after standardization, will be driven by shared tenancy models and this deal positions us to take advantage of this model in the medium term."
We think he means that it's cheap.
A News International spokeswoman said suggestions that the outsourcing programme was to blame for the outage were "not correct", but refused to discuss the cause further or make any other comment. ®
Watch Now : Virtual Machine Movement with Hyper-V
COMMENTS
It's possible...
That the reason he's not at Accenture anymore is because of stuff like this.
He's right, though, it is entirely possible to run a lot of these systems with off the shelf stuff and low skill staff.
As long as you don't mind the frequent outages and service problems created by using something that's not fit for purpose.
not in my experience
I work with HCL, and in every case our inhouse team have had to do more work supporting the outsourced work, than we would have had to do to build the thing ourselves.
It's not about skills, as we've had to train the offshore teams. The only way it's cheaper, is on paper. in that you get a project manager, and offshore coordinator, and onshore coordinator and a developer, for the price of one developer. The actual productivity is less.
Somebody somewhere gets to say i have a team of 200 people offshore, for the same price as 20 inhouse developers, no-one ever ties back that you probably could do the same work, to a higher standard with the 20 inhouse developers.
I continue to maintain, that offshore outsourcing, just isn't cost effective. You have the language barrier - yes everyone speaks english, but not fluently. pidgin English code comments and design documents are a nightmare to unpick. Also, your carefully worded spec, may or may not be understood in the same way as you intended.
You have cultural differences, ties back to language, but this is a biggie, this is one we rub up against constantly, it manifests to us as no-one using any initiative whatsoever, you get exactly what you ask for without question or consideration that it will do what it's supposed to.
Things work differently within companies too, we never seem to work with the same group of peopel twice, so every project is a training excercise on our part. We train up a team of developers, just get the code up to a standard we can accept and the project ends, next project, rinse and repeat with a new team.
You have the simple timezone differences, which means a 5 min email back and forth can take a week due to people working at different times. Having no physical presence is also more problematic than you'd expect in IT, being able to get everyone sat around a table together, to go through designs is invaluable. Sharepoint isn't a substitute!
One of the problems we've had, is that we aren't allowed to tell the people higher up the problems we've had. It can be a train wreck from start to finish, we step in and pull a project out of the gutter at the last minute, as we'd take the fall if it failed, and that gets reported through as "<offshorer of choice> did a great job in getting that last project implemented, lets use them again for the next one"
You said it....
I worked for News International for 10 years until Thickey made me surplus to requirements. (Thanks for the fat pay-off by the way Andy). During that time News never had an outage this serious, mainly due to the calibre of staff that were employed during that time.
Furthermore, they never lost an edition or came close to it.. Now as the Indomitable Gall alludes to, the true measure that outsourcing works for the Newspaper industry is whether the paper comes out during the 4 hour or so daily production window day in day out.
General opinion of the remaining IT staff is incredulity that they have managed this long without losing an edition, given the incompetence of HCL (or to be fairer, their ignorance of Newspaper production) It's fair to say it's not if this will happen but when. Then will the shit truly hit the fan.
They have even managed to lose the source code to their in-house written Editorial Production system for God's sake. Bet the auditors would love to hear about that...
A not too disgruntled ex-employee
Off the shelf outsourcing never works.
You can see this happening anywhere.
It's the same old cycle.
Innovators create the software.
Completer finishers finalise it.
Architects (now the product's making money to afford them,) come in, and talk about models all day, introducing a few third party products of their own, "because the existing solution doesn't scale" despite it running on sub 1% resources, and outselling all the company's best other assets.
Architects having caused a few problems, "software engineering" consultants now come in, usually sandlewearing beared types who talk with a nasal tone, know the W3C standards by heart, and who've read software engineering from a book, and never delivered anything, and say, "You need some design patterns, N-tier, messaging and outsourced search." to gullible management who pay them a grand a day.
This fails, because it's one size fits all bollocks and the consultants have never delivered anything. The scope is changed from faster, and more scalable, to "fast enough" and "it handles all the customers it needs." by middle management to avoid having to say "We just spent hundreds of thousands to get to a worse place."
The management of the day, hire in more children, plus one guy in a £2000 suit, in the form of the bane of techie's lives - management consultants.
The children go around with clip boards interviewing the staff, and the man with a £2000 suit makes a report saying the problems are what they've just been told and he "recommends an off the shelf" product because there's no work for him to do in this case.
All the techies that can leave, being disgusted, leave, usually including the Architect, who moves onto another architect role, and all the steady talent that hung over from the original setup.
The company is now at the mercy of management consultants who bleed the place dry, because the upper management value the statements from a man in a suit, over the statements from the Autistic IT guy in stained polo shirt and paint spattered (at least it looks like paint) track suit bottoms.
Tzimples.
Or in other words
"He believes all IT departments from banks to supermarkets to newspapers can be run in the same way: ie just using off-the-shelf software administered by low-skill employees,"
He's a fuckwit!
The new Office Garage series:
Enabling efficient data center monitoring
IT infrastructure monitoring strategies
