I’ve been speaking to some folks about Windows 10 lately. I ask about it a lot at conferences, too. The general feeling, personally, is that about half the world is deploying, or gonna deploy. The other half, “meh.”
You know why the meh? Everyone probably wishes they had Windows 10 deployed, but almost nobody is excited about deploying it. It’s kind like me and “shoulders day” at the gym. Love having done it, not so hot on doing it.
Close your eyes and go back to that happy, happy place and time, when your users all ran Windows 95. Then, Windows 98. Probably that was it. Lots of companies never broadly deployed Windows 2000 Workstation – it was a massive change over the MS-DOS based Windows versions, and they just weren’t up for it.
But along came Windows XP.
And it stayed. So not only most IT Ops people, but most entire companies, have only ever deployed a client OS maybe three times, max. Once, when you brought in Windows 95, which you probably did on new computers anyway. Again when you deployed Windows XP, and again when you finally rolled out Windows 7. Yes, lots of folks have done more, especially if you go way back, but the mass of the world seems to have done about two or three.
And it’s largely because deploying new client operating systems is a pain in the ass. It’s literally the most impactful thing IT can do to a company, because bloody everybody sees it. Everyone gets a new Start menu, a new Control Panel, and a new reason to bitch at us. Migrations never go 100% smoothly, because users stash stuff everywhere they’re not supposed to. Migration tools suck in large part because we don’t use them enough to generate any interest in fixing them.
And migrations are mainly a PITA because nobody has invested in building the skills, tools, or infrastructure to make it easy. Yes, Microsoft has released tools – lots of them, over and over – but none of them are a Magic Migration Deployment Wizard. Because migrations have always been this painful-point-in-time thing, we’ve never made it an ongoing process that requires us to make it a science. When we deploy OS v2, we don’t put in places the things to make the eventual OS v3 deployment easier. When we acquire client apps, we don’t bother even asking what the forward compatibility story will be.
And so client deployment is a mess.
Windows 10 might offer some relief. Maybe. It depends a lot on how long Microsoft follows through with this servicing branch theory they’ve got running, where most of our computers perhaps live int he Long-Term Servicing Branch (LTSB), but eventually suck down whatever updates are sent to that branch, while others can opt into a more-frequent series of updates as a means of testing new features and fixes. And if they stick with this as a means of shipping new features, eventually shifting to a subscription model rather than a one-time-sale model for client licenses. That’s a lot of “ifs” from a company that doesn’t have a great track record for sticking with update strategies (“we don’t ship new features in service packs,” lolz).
So either Microsoft takes this off our hands by making clients inherently “always upgrading,” without the need for backbreaking deployments and migrations, or we have to fix this ourselves. “Fixing” meaning actually building up the tools and processes so that if someone says, “yea verily, go thee and migrate the Sales department to Windows 12,” we can just nudge a button and have it happen. These damn client computers are our single biggest source of maintenance, upkeep, and security woes, and we need to be able to keep them updated – beyond just patching them.
Oh, also – I committed T-Mobile.