IT infrastructure on demand? Yeah right, say devs

IT operations remain completely out of touch with the needs of developers, with CIOs duped into believing a dusting of VMware magic will allow them to construct the sort of whitebox data factories that power the likes of Google, sandbox vendor QualiSystems has declared. The vendor’s CTO Joan Wrabetz took aim at the industry’s …

[Anonymous Coward]

I don't really believe any reasonably sized company and beyond have all the things required.

1) The right staff

2) The right number of staff

3) The correct physical infrastructure for their applications

4) The correct application delivery estate

5) Vendors and Developers that work in partnership as opposed to a kleptocratic nightmare

6) Enough time to do things right

7) A continuous process to improve on the estate

8) The correct management and monitoring suites for their estate

9) Succession planning

[Halfmad]

10) Continual, reliable investment to make it work for more than 12 months if 1-9 are in place.

[ro55mo]

Oh, I'm sorry

That I don't have boat loads of unused capacity ready to be handed out to users when unplanned requests for resources splat onto my desk.

[Steve Davies 3]

Re: Oh, I'm sorry

But... that's what the 'Cloud' is for. Time to get on message then...

[Anonymous Coward]

developers can be twiddling their thumbs for more than a month

Corrected, developers/IS can be arguing over specs for more than a month.

Anon because I don't fancy waiting more than a month for a bog standard server that we then actually have to configure for services/software.

[Franco]

Translation: We sell a product and are in need of a niche to market it to.

Sales pitch masquerading as industry analysis.

[Adam 52]

Yup. And his research seems to disagree with Gartner snd Forrester, wonder why that could be?

[John 104]

Joke Alert?

Right?

I find and have found, in 16 year of being in IT, that devs really have no idea what hardware is, how it works, how long it takes to spec, req, and acquire, etc. Cloud based infra helps speed things along to a degree, but it doesn't eliminate the problem of requirements vs reality.

As for users, they are probably about the same as Devs. Maybe Devs are worse as I expect them to have better knowledge of the systems their software is designed to run on...

And, to ro55mo's point. VMWare isn't free. Hyper-V isn't free. You don't just get to spin up an OS and forget about it. You have to PAY for it just the same. Same with physical assets. Management doesn't want to spend the money on good equipment for users or the server room, so IT is left holding the bag when the old stuff breaks.

[AbstPoolAuto]

What he said ^^^^^^

Been a couple of naff articles like this over the last couple of days. The usually good Chris Mellor had a similar one on Netapp earliar - bias masked as analysis.

[Dominion]

FTFY

"Developers remain completely out of touch with the needs of IT operations"

Is actually closer to the truth. Time and time again we see solutions that cannot be deployed to production environments because the slap dash "sandpit" servers that were used for development is using OS builds that are not representative of a properly secured and hardened production server. If IT operations are "utterly out of touch with what developers, users and businesses actually need", it's because developers, users and businesses don't have a plan or a strategy to enable their work to be delivered, and didn't think about flagging up their requirements in a timely manner.

[Roland6]

Re: FTFY

As an IT Architect, I ask myself, to what extent does the typical IT operations department need to be in touch with developers and who's fault is it really if they aren't?

I've found that many IT departments use systems integrators to design and build new stuff that gets delivered and handed over to IP operations, with the systems integrators (typically external consultants) acting as the go-between between in-house and third party developers, the business and it's users and IT operations.

One of the biggest problems I've repeatedly had is that IT Development are the worst users for knowing what it is they actually want and need, which can cause problems when you need to keep the team in India in sync with the UK team. On one project the dev lead thought they would achieve this by him taking an image of his HDD and all of it's test VM's and then FTP it across the wire - he failed to understand until he had burnt a set of DVD's that an image of a well used HDD doesn't compress very much and we weren't going to attempt to send a dozen DVD's worth of data across the Internet every night...

Hence designing a system to support development that can be delivered to IT operations becomes highly problematic.

[Ilsa Loving]

It varies

I can see both sides of the story here. In general, yes, even if the IT dept is working perfectly, you will still have issues because as others have said, a Dev (or whoever) still has an excellent chance of making an unreasonable request, if for no other reason than they simply don't know what the limits are. A good IT shop will fish out the details so that something viable can be done.

I've also seen IT that I swear raise Mordac (the Preventer of IT Services) to godhood. It's awe-inspiring how some people can stonewall a process.

That being said, the part of the article complaining about how it can take UP TO A DAY to provide a user with what they need, and I just scoffed audibly. The last thing we need is someone who thinks Visual Basic is the pinacle of programming, to dictate infrastructure.

[Doctor Syntax]

If your business thinks that 8 hours or even a week is reasonable notice you're doing things wrong. If you have some project that requires resources then the time to bring operation into the planning is at the start along with everyone else.

Having said that I have been in the situation of arriving at work one morning and being grabbed by the data input team because they had several temps just arrived to enter data for which no software existed because it hadn't even been mentioned to anybody in IT. And another instance when my first meeting with the project team was a day or so before go-live - fortunately I'd heard about that one on the grape-vine & although I hadn't done any coding I'd thought it through & realised it wasn't difficult. Those were exceptions because usually I and my team had good working relations with the rest of the business and because, in those far off days, we did S/W, DBA & Unix admin as a single team we didn't get into the sort of face-offs implied by this report.

[Anonymous Coward]

Damm new kids on the block

Don't have a clue. End of, full stop.

I had to stop on from Ops last week from taking a node out of an operational production MSCS cluster without getting an agreed outage or a backup plan if it all went pear shaped. Then he had the nerve to complain that said outage was from 02:00-->04:00 on a Sunday morning.

The real world dictates this sort of thing not some head in the clouds lecturer at University.

So tomorrow morning I'll be doing the job because the 'new kid' is gonna be out partying/clubbing and most certainly not in a fit state to do the job in hand.

Us greybeards are retiring. We have 40 years of knowledge and experiece behind us. I started when we toggled bootstraps into front panels and did backups on 2400ft magtapes and disk capacities were measured in tens of Mb. We will take that experience with us. I'm calling it quits in June. God help many companies when we are gone. And no, I'm not going back as a contractor. When I'm done, I'm done.