While I don’t think we’ve gotten to a nailed-down understanding of cloud computing, there is one way to slice that is interesting, and often overlooking: dividing up the types of cloud computing users, primarily along the lines of ISVs and IT departments.
If “cloud computing” means infrastructure outsourcing, keep your eye on the ISV/IT department fork
What’s clear is that the point of cloud computing is moving your software mostly out of your control and depending on your cloud provider to actually run it.
Now, the effects of that instantly split along the ISV/IT Department axis. As with any discussion of infrastructure, this is the first axis you have to splice on, or things get weird.
Pro-tip: your first question to all cloud providers should always be “is this for ISVs or IT Departments?”
There are several, we theorize, benefits for each. Far from an exhaustive list, the next two sections go over some low-hanging positives:
Benefits for ISVs
If you’re an ISV – someone writing software to sell to others, not use yourself – you’re going to run your software on a cloud, using the SaaS business model. (Throw in re-sellers and The Channel here a la Jamcracker, OpSource, &co. for some angels on a needle-head nuance.)
Here, you’re building a SaaS. So, of course, you get all the benefits of being a SaaS ISV (distribution, volume-pricing, easier maintenance of one instance instead of n+1/customers, etc.). The benefits of cloud computing, though, are based on you not having to manage the infrastructure that runs your SaaS, for example:
- Boot-strapping costs are reduced – you don’t have to buy those data-centers up-front. This means it’s cheaper to fail, which means you can try out more ideas more quickly to find the one that’ll be expensive enough to make money with. It seems like ongoing costs will be the same, if not more if you use only cloud computing. Another Velocity panel – “Surviving Scale” – hit on this: all of the big-success panelests had their own data-centers and just used The Cloud as supplementarity computing. Force.com, of course, is big into this boot-strapping angle, as are it’s PaaS buddies, like Bungee Labs.
- Scaling will happen when needed – instead of waiting to get failwhaled, your cloud provider will (magically) make it work. At Velocity, the network performance and optimized guys ruled the roost here. Our CDN friends should (we hope) be readying their swooping-tropps here in the same vein as people like Rackspace were smart enough to shell out innovation here from the mother-ship to Mosso.
- Finally, an SOA We Can Get Behind – if you throw in the idea of middleware as part of the cloud – see things like Microsoft Astoria and ESBs in the sky – you start to get to that web-like, URL-bound programming we’ve been lusting after for years. Sure, it might be a lot of XMPP instead of HTTP, but who’s peeling back that low? (Oh, everyone. Right.)
- Google Might Acquire You – if your cloud provider is into M&A, they’ll be on the look-out for people to acquire. IBM and Microsoft do this in the on-premise world – buy successful third-party ISVs that run on their platforms/ecosystems – and there’s little doubt that it’d be wise for cloud providers to shift into the infrastructure & applications market that’s worked out well over these decades.
Here’s one thing very few cloud providers are currently promising developers: seamless transition of skills and code. Indeed, most cloud providers are (relative to the anarchy of on-premise/in-house SaaS development) hideously proprietary and their platform.
To be fair, no infrastructure paradigm is ever seamless. Indeed, I’d suggest that the idea of abstracting away your underlying infrastructure – what your software runs on – is fallacy. Sure, we can all agree on a shared-madness – J(2)EE, .Net, Spring, OSGi, rails, RIA’s, etc. – but you, dear code-monkeys, will have to re-tool.
People like RightScale have some interesting normalization going on here, but it’s still their own, not ours.
Benefits for IT Departments
If you’re an IT Department (“enterprise,” if you prefer), you’re looking to move the applications you run for the business into the cloud, and move those money-makers out of your own data-centers.
As with ISVs, the benefits come from you not having to own and run the infrastructure that the applications run on, for example:
- Staff “Optimization” – you’ll either lay-off or re-assign IT staff (in- or out-sourced) to save cash or re-focus on “more valuable” uses of their time. At this point, the boxes-and-arrows set comes parachuting in suggesting we can replace Eclipse with Visio. Over the past few years, 3Tera, it seems, has managed to move itself out of being lambasted on that account – good on them. But beware people who tell you the IT department can be reduced to right-clicking.
- Cloud Arbitrage – if you can get some capacity management wizards and the standards making it possible to quickly move from cloud to cloud, you can forever be on the look-out for the cheapest cloud provider. Cloud providers haven’t cottoned onto locking users into long-term contracts (yet?), seeming to charge mostly monthly subscription fees, at least when it comes to list price. At this point, this is pure utopic thinking, but IT Departments would be wise to start pressuring providers into this direction. One word here: standards.
- Scaling Out – as with ISVs, one of the top promises of cloud computing is the elastic nature of it: as you need more capacity, it will be added cluster/grid/etc. like. This of course, assumes that your applications support that, which is a big assumption. Still, you have to think this is better than taking order on truck-loads of blades a week.
- Avoiding Optimization – in theory, if someone else is running your infrastructure, you can avoid having to be and pay for being an optimization wizard. At worse, you’ll need some ace network performance management dudes. This frees up time for…
- Focusing on Business and Supporting Process – the pay-off any IT Department want is to show The Business how all this IT is directly making money for The Business. All else is beneficial distraction to that point – yes, we realize it’s necessary, but The Business doesn’t tolerate such reality. If you can actually outsource running your infrastructure to the cloud you can (Lord, we hope) remove much of that distraction and spend more time proving to The Business why the IT Department should continue to get funded. (Psst…BSM vendors, get it rolling here quick!)
(Thanks to The Onion for my title template.)
Disclaimer: Microsoft, IBM, and RedHat are clients.