"In our country," Alice told the Red Queen in Through the Looking-Glass, "you'd generally get to somewhere else – if you run very fast for a long time, as we've been doing." "A slow sort of country!" the Queen replies. "Now, here, you see, it takes all the running you can do, to keep in the same place. If you want to get …
It's hard to get a sensible amount of money for IT spend.
We don't know how to get more.
Dunno what to do.
Oh, I'll write an article full of management speak. At least I can get money for doing that.
Hard problems are hard. Who knew?
Management & Sysadmins
IT hardware is expensive, good sysadmins are expensive and a poorly functioning IT infrastructure is very expensive.
In my experience sysadmins aren't especially greedy. With this in mind, I have never understood, why sysadmins are given budgets. The sysadmins should be able to say what they'll need, when (and why, of course) and signed cheques should follow.
Management wants consistency in its IT-infrastructure. IT should be almost invisible and only noticed when its not available. It should allow the company to run smoothly.
In my experience, it is upper management that decides what technologies get used, and, as a rule, it tends to be something expensive like Oracle, anything from Microsoft and the like. As a rule, sysadmins would choose something different, something safer and less disruptive. Furthermore, too much interference by management in the running of IT-systems drives the good sysadmins away and prevents an optimally run IT-services.
In short, my point is that management should look for good sysadmins and let them make the technical decisions. One doesn't go to a doctor and tell him the way you want your condition treated. One doesn't go to a programmer to design a database. One doesn't go to a marketing specialist for decisions on SANs, choice of OS or server hardware.
And another thing...
Excellent article for those of us in large IT shops.
Another way to minimise risk is to take your weakest component process wise, develop and enormous cumbersome release management process and then apply it equally to all systems making change near impossible. Update it everytime something goes wrong in any change, no matter how irrelevant, and increase the paperwork burden. Pretty soon the only change is emergency and critical patching and IT spends all its time filling in forms. Job done, budget used.
I blame ITSM/ITIL for that one.
For the next one, I blame techies... software defined everything = an enormous increase in workload. There was a time that a switch would get a firmware update once a year. Now it gets at least a monthly security patch in addition to fixes for the inevitable outage causing bugs. The same for DASD. The same for servers. For a stable workload (flexibility not required), the cost of software defined everything must be well in excess of the benefits.
Blame the Red Queen
You can't talk about Larry like that!!
Then there are the gravediggers
Who have to pick up those abandoned systems, tear them to pieces, toss the bits into the correct updated piles, and then put the now dead creatures out to pasture, usually on a forklift to the dump.
Can someone tell me please, why it is that when you've been herding the snowflakes that management insist have to exist, you have to shovel management out of the drifts later?
Did he say anything? I vaguely remember some sort of contrived simile and then I nodded off.
As yoganmahew said, good article for those of us in large IT shops
I think the people missing the point of the article are probably working in smaller shops with newer shinier toys. I know some people today desperately patching Windows Server 2003 and XP machines. I found this to be a good article. Not for myself necessarily but I'll be sharing it with some people above me who *do* get to determine my budget.
Re: As yoganmahew said, good article for those of us in large IT shops
Is it the same shop where someone is desperately looking for an owner for Win Serv 2003 machines so they know if they can be restarted to patch?
The root of this particular evil is that spending decisions are made quite high up the management chain... which means that all too often money gets spent on high-visibility projects that are career-enhancing.
And once the shiny new system is in place, it cannot be seen to fail, or its sponsors would lose face.
So a big spend on new shiny can certainly such the life out of maintenance of legacy systems, especially when they are unsexy and wouldn't contribute to the sysadmin's CV.
If only it was possible to kill the belief that transformative system upgrades are still possible, the industry could quietly get on with step-wise upgrading based on a long-term system renewal plan.
Indeed when a new system is commissioned, the hand-over should include a succession plan laying out the upgrade cycle and ultimate end-of-life for that system.
talk about risk
There is a lot of talk about "risk of change" but IT does a very poor job of explaining to the business "risk of NOT-change". Just because something is stable now doesn't mean it will stay that way, and as things get older the chance of something going wrong is only going to increase - especially as other systems get updated.