I don't buy this article. Nothing is "hindering" cloud adoption, per se; the question is, what problem does cloud adoption solve? For me, the answer is "relatively few:" I have Mail.app for e-mail, Word / Mellel for word processing, NetNewsWire for RSS feeds, Textmate for text editing / blog posts, and so on. None of these will be particularly improved by cloud computing.
But some of them will be harmed by cloud computing: I don't want to have to have Internet access to access my files; I don't want other people to have access to my files; I don't want to have to learn a whole new set of tools that aren't improvements on my old set of tools.
In short, the benefits outweigh the costs. When/if the benefits exceed the cost, I'll start using the "cloud" more. The advocates of "cloud" computing aren't empathizing with their users.
Personally, I imagine that we're heading more towards a place where we have more Dropbox-style local / network integration.
Reading that it seemed as if OM considers it a given that cloud computing is automatically "better" and squishy humans are just getting in the way of progress.
Cloud computing is in many ways more efficient, but that's really not the only metric that matters, is it?
There is value in independence, privacy and security - none of which cloud computing can ever really compete on. There are also many applications where proximity of data or proximity of processing matters, and the cloud can't touch those either.
There's also cost. At certain scales (such as Grooveshark's scale) it can be significantly cheaper to do infrastructure in-house. Try negotiating bandwidth prices with Amazon.
Actually, I talked with them at openworld a while back, and i got the impression that the prices were negotiable if you're a big enough customer. I may have totally misunderstood them though, but I also talked with them back when they were significantly smaller (though growing rapidly).
the big advantage the cloud brings is speed of spinning up new servers, for which it's nearly unbeatable. If you need 1000 servers tomorrow and want to get rid of them on Friday, there really is no other option. However, if you can forgo that speed, if it's okay to take some time to get the servers up and if you plan on keeping them online for a few months, owning starts to make sense around the time you need 16-32GiB ram, which really isn't very much.
Agreed, however the article is about adoption of cloud services. Hopping on temporarily for extra capacity doesn't seem like the kind of adoption they are trying to encourage.
The other problem is that if you run a mix of in house and out of house services, the latency between the two sets of servers may be too high to be of much use. Say for example your DB server is on your network, and you need more capacity on your web servers. You could fire up a few extra EC2 instances and set up a VPN link from them to your DB server, but most likely the latencies are going to be too high to be useful. So cloud services aren't necessarily all that useful for temporary capacity either. To some extent it's all or nothing, at least for web services.
You're assuming either/or. I have a mostly in-house IT infrastructure but I use EC2 not only for projects that I need to get up and going quickly but also for traffic spikes. Using a dynamic DNS like Nettica allows you to blend quite easily.
I think that a mixed setup like you describe can allow you to get the best of both worlds, if all your servers don't need to be in the same place for latency reasons. "the cloud" is also awesome for backups. you can have a full backup cold spare system, test it once a week, and pay almost nothing for it. (this is what those 'micro' instances are for. you have a micro instance updating your DB on an EBS backend. when it comes time to fail over, shut down the micro instance and bring up the ebs mount on something beefy.)
my point was just that most systems are on most of the time, and for those systems, owning saves you a lot of money over using "the cloud"
yeah, if you only need a few hours a day, "the cloud" is the place to be.
But, most workloads are on 50% or more of the time, (see the popularity of amazon's reserved instances) and you save a lot more than 50% owning your own hardware, unless you are paying ridiculous fees for your hardware guy.
But some of them will be harmed by cloud computing: I don't want to have to have Internet access to access my files; I don't want other people to have access to my files; I don't want to have to learn a whole new set of tools that aren't improvements on my old set of tools.
In short, the benefits outweigh the costs. When/if the benefits exceed the cost, I'll start using the "cloud" more. The advocates of "cloud" computing aren't empathizing with their users.
Personally, I imagine that we're heading more towards a place where we have more Dropbox-style local / network integration.