Alternative IT Thinking

My previous post on "frictionless IT" kicked up a bit of dust, if the email traffic is any indicator.

I think I ended up carving an interesting line between what IT has traditionally been, and what it's more likely to be in the future.

Clearly, not all of IT can end up in a self-service, few-questions-asked environment -- at least, not anytime soon.

But I think it's a fair question as to where we'll start to see newer forms of "utility-based" IT models spring up around virtualization -- and whether the utility model will actually extend to the users.

So, What's This All About?

Previously, I argued that -- given the advent of technologies such as virtualization, Web 2.0 and more -- perhaps leaders in IT might want to start thinking in terms of "frictionless IT" -- powerful services made available to users on demand with as few barriers as possible to their use.

I gave examples around test and dev environments, business intelligence, and a few others -- situations where it might be true that technology has tipped the balance to the point where simply putting users in charge might be more productive than having IT engage in a traditional manner.

Some Common Examples Today

If you think about it, many of us have access to IT services that are essentially frictionless -- we don't really have to go through too much process to use them.

Email is a good example. It's just assumed that everyone needs access to email, and -- generally speaking -- the benefits of making it free for all outweigh any processes associated with justifying its use.

I could say the same for internet access -- although getting to certain sites that someone somewhere has deemed as "bad" can be a bit problematic, although that's changing.

File server space is another example. I don't have to justify to anyone as to why I might need 20GB or so on a backed-up corporate file server.

Here at EMC, we've got several collaboration environments that don't require much in the way of justification to use: EMC's own eRoom, Microsoft's SharePoint and Jive Software's ClearSpace are all examples. Not to mention con call birdges, WebEx, et. al. It's just assumed that you're gonna need some of this stuff to get your job done.

Now, to be fair, for each of these environment, someone has to be monitoring the environment to make sure that it's available for everyone, it's being run efficiently, the use of the platform isn't exposing the company to risk.

And, of course, when any these platforms get large, they create all sorts of follow-on opportunities for optimization, management, etc.

But -- just to be clear, there's little scrutiny around who's using it, and what they might be using it for. It's just assumed that people need this stuff to do their job.

So, Here's The Question ...

At what point does it make sense for IT to think in terms of a self-service model for a given IT resource, as opposed to a case-by-case or, perhaps, a chargeback-for-use basis?

Incremental cost is definitely is a factor. Sure, up-front infrastructure costs are expensive -- and have to be covered somehow -- but, after that, the incremental cost for the next email user, file system user, etc. is quite reasonable.

So, given that you've got a nice virtual server farm, and everything is decently automated, what's the incremental cost for, say, a Linux application stack that's only occasionally used? Not that much, I'd argue.

For our social productivity platform here at EMC, anyone can jump on, join a discussion, start a community, post stuff, etc.

We know exactly what the incremental cost is per additional user. It's not a lot ... and we'd rather have a new, contributing participant rather than fret over whether a given individual should or should not participate. It's not worth the time to evaluate or discuss.

Let's say that you've exposed multiple data sources (either realtime or static) in your environment, and have provided some nice analytical and/or mashup tools to your users to experiment with.

What's the cost of an incremental user who creates something that might be of some use? Relatively small, I'd argue.

Utility Computing Redux

I can't tell you how many "utility computing" discussions I've been in over the years. And, just about every IT shop I talk to has this mental picture of a humming VMware farm, very optimized and very flexible, almost running as a server farm utility.

But these same people seem to insist on a model where they're still the middleman. No one gets on the utility grid without a standard propose, review and chargeback model.

I don't know if I'm making any friends when I suggest that acting as a traditional middleman might not be the best answer going forward.

Why not expose the utility directly to the business? Or, alternatively, how much value is generated by IT acting as a gatekeeper?

Building The IT Nation Of Tomorrow?

Another analogy that strikes me is the role of government, especially here in the US.

We all pay taxes, but -- at the same time -- we have the potential to use an never-ending array of "free" services, which include education, public safety, research, defense, and so on.

Sure, we can argue about the quality of the services, and the efficiency with which they're provided, but I don't hear many people seriously arguing that we ought to go back to a pay-as-you-go philosophy for most of our public services.

I, for one, though, am hopeful that corporate IT groups can do a far better job at providing shared services than our elected officials ;-)

Consider The Big Trends

IT inputs are getting ever-cheaper: compute, storage, network -- even software, once you start considering open source. What was once scarce and expensive is becoming more cheap and plentiful with every passing month.

At the same time, our workforce is changing. We're becoming an economy of knowledge workers -- we deal in ideas, expertise, analysis and communication. Few of us are the transactional workers of yesterday.

Our newest workers have grown up in a world of technology privilege. They've always had personal computers, the internet, mobile devices and so on. By and large, they were always "just there" -- they never really had to justify what they wanted to do, or ask permission.

And, in our never-ending quest for IT efficiency, what IT services will be likely candidates for moving from a justify-what-you-need, pay-for-what-you use approach - to the ultimate "IT utility" where generic services are provided at the lowest possible cost, and it just doesn't make sense to scrutinize each and every user?

Once Again, Virtualization Changes Everything

Given the interesting discussions I'm having with more than a few people, I'm guessing that the widespread advent of server and desktop utilization will drive yet another batch of IT use cases into this self-service model.

It's easy to see that -- before long -- it'd be easier to provide a developer, or an analyst, or any knowledge worker with a no-questions-asked stack of compute resources.

It helps if you don't think "server", and think "server image" -- actually a file that gets loaded when it needs to be used, really not all that much different than me loading up a big PowerPoint file on my desktop.

Sure, There'll Be Some Work To Do

Just like with other past examples of shared IT services, there'll need to be someone watching out for the environment, making sure it's available for everyone and not consumed by a few "proficient" users.

And, just like file servers, email, et. al. there'll be a never-ending sequence of after-the-proliferation "efficiency enhancements" to these self-service compute environments.

But, for a large part of our value-generating workforce, much of the friction associated with accessing and using information and compute resources will be removed.

And the resulting leap in productivity and business benefit should far outweigh the "inefficiencies" associated with a shared services model.

It's happened before. It could happen again very easily.

What do you think?