Re: Time to call "Time ladies and gentlemen" on the cloud...
+1 to your "passing the buck" comments Dan.
You could pretty much argue that hosting your servers in a rack in a DC is also "passing the buck", because what happens when the DC provider (or your link to them) goes down? So now you're building your own DC-grade facility at your site? How's that working out for you? Nice and cheap to do properly?
The further you move up the sophistication chain, you're trading off nuts-and-bolts access to the server for something even better - SLAs. Make sure your SLAs with your providers are solid and business-appropriate. That way, when things DO hit the fan someone else has to fix the problem for you. When faced with the choice between staying up all night sorting a borked Exchange IS or making a phone call and having my provider do it for me, I'm going to choose the latter.
(Valid point to be made here about how sophisticated you need to be. If you consult to small businesses, they may be happy to use a less reliable system based on consumer gear if it keeps the cost down. I'd still argue however that there's benefits to be had by cloud-ifying some of their stuff, even if you start with something like Dropbox for Business that's less integrated but more affordable than an enterprise solution. Horses for courses.)
And I can't comment for other countries, but I've got to admit - our data links (be they private IP VPN links or basic internet carriage) are stupid-reliable these days. I honestly couldn't tell you the last time we had a link outage. And that's going right back to frame-relay days. Even the one 3G link that we're using is reliable (albeit as slow as a wet week, but it's better than a $2m fibre build out "investment").
Look - Office365 still has some hairs on it. We've been working on hybrid Exchange Online for the past year, and have just decided to pull back from it because there's one specific piece of functionality that we need (business-wise, not tech-wise) that we simply cannot do. So for now, we need the extra power and flexibility that you get on-prem. But you can bet your bottom dollar that the moment MSFT announce a solution for us, I'll start migrating mailboxes.
OneDrive FB is a dogs breakfast (all they had to do was copy DropBox, and they buggered it up...), and I'll agree with the article on Sharepoint Hybrid - you can do it, but man, you've GOT to be committed (or you will be, after you spend a bit of time with it). That said, we're utilising SPO for a variety of sites that integrate seamlessly with our on-prem SP without having a proper hybrid setup. The biggest UX hurdle is seamless authentication, and you can solve that pretty easily via IDp-initiated-signon and smartlinks. The rest is just look-and-feel.
Lync Online has been a real success story for us - click button, get enterprise UC/IM/video system. We'll be moving it back on-prem soon, but not because of any problems with it. We want to do Enterprise Voice, and there's been only the vaguest murmurs from MSFT about when that might be available. Moving on-prem, we can do it ourselves in 6 months. As above though, you better believe that as soon as Lync Online has robust EV capability I'll be migrating back and divesting myself of having to run what is a delicate and expensive backend.
The article has the right of it, and I'll also back up the other commentards here. There are some things that the cloud is excellent for (reference the comment made about hosting your website internally - that's gold), and some things that you need on-prem systems for.
The masterstroke of the MSFT approach is when you pull Azure into the picture - then you get the ability to run on-prem installs of systems inside a cloud DC that's an extension of your internal network. You get all the beauty of cloud management (easy resource assignment, scale up/down, op-ex payment structure, etc) with all the customisation potential of an on-prem deployment. And as someone who has spent the last 6 months facing extremely constrained datacentre resource capacity with no easy solution other than large cap-ex, that's a very attractive option.
My $0.02 - I'll caution you that it's probably worth markedly less than what you paid for it.