Self Service - IT Nirvana for the Enterprise?

It’s been an exciting few months at Cloudy Towers, and I’ve let my blog writing lapse somewhat while we’ve been building out new areas of focus and gearing ourselves up to deliver even more awesomeness over the next year.

Fear not, I’m now back in the game. Following a number of interesting conversations with large global clients this month, I thought it worth revisiting a topic I’ve touched on previously: Self Service IT.

Remind me, what?

If memory serves, I first mentioned this in December last year, and I’ll pick a quote from that blog as a quick refresher as to what I’m talking about:

"Even if you were to only pick one cloud provider, you would almost certainly want to implement a degree of self service, whereby teams and individuals with appropriate permissions can request, and thereby programmatically create, infrastructure via a portal of one form or another. If you have multiple cloud providers, you will definitely want to do this to provide a layer of abstraction. Your end users should only care about workloads, not underlying infrastructure and which platform is being used."

So, we’re talking about allowing people to request "IT" on demand, either infrastructure, or ideally infrastructure and pre-configured and deployed applications.

Within large organisations - think Fortune500 type scale - we’re definitely seeing "self service" as a core trend now. It’s worth noting that most are still at the "discussion" stage, rather than implementation.

People are really doing this?


Earlier last month AWS formally released Service Catalog (or Catalogue for the confused Brits needing a translation). It’s been in trialled by major clients since release in late 2014, and has now graduated to a full service following feedback and enhancements. It’s actually still free, yes free, until the end of September, then it defaults to a nominal charge ($5 per workload per month).

I’m expecting this to drive a lot more activity in this area for upper-midmarket sized businesses during 2016. The downside for the larger enterprise is of course the inability (currently) to launch services in other cloud platforms.

He’s talking about 'multi-cloud' again

That’s true, I am.

Note: I’m still a passionate believer in "backing your horse" and just getting on with it, i.e. evaluate the cloud offerings available, pick one, and realise the benefits as soon as possible, not in months or years.

In reality, it’s often not realistic for larger enterprises who for various risk and compliance reasons will want to choose multiple vendors. Equally, there may be general good reason for opting for more than one public cloud vendor - I had a stimulating conversation with a Gartner analyst in the week who coined the phrase:

there is no 'cloud', only 'cloud services'

By that, he meant that one should pick the best bits of the different providers according to your workload needs. We’re certainly seeing more of that activity now, and it’s something perhaps to be applauded - so long as you’re comfortable that it will take longer to establish the processes and mechanisms to do this at scale.

If you can do it though, you’ve essentially managed to achieve the holy grail of IT. Imagine a world where a team can simply request a new workload via a portal and have it automatically provisioned on the most appropriate platform available in minutes - with no interaction with "IT". Yes, they’ll customise the deployment to their own needs thereafter, but you’ll have provided a workload with all security, compliance, software and architecture versions and rules applied by default andstandardised.

Sounds easy, let’s do it

Well, as is often the case, the best things in life require a concerted effort to achieve:

1) I’ve found myself repeating this a bit recently: "It’s not an IT project, it’s a change project." In reality here, it’s both - but if you don’t manage the change, this concept is going nowhere. Your local business units, project teams or group companies will almost certainly not want to be "governed" by a central cloud broker - even if it will make their lives easier. What did the Romans ever do for us? (second Python ref. in one blog…)

You’ll need to master politics as well as IT to make this work.

2) Politics aside, there are interesting accounting challenges to make this work at scale so that your business units feel they are getting the best value for money. You can be much more granular and specific with regards to apportioning cost. Which is either a good thing or a bad thing depending on how it affects individual budget owners...

3) There is a lot to consider. There really is. In a multi-cloud, or a multi-service world, tooling exists, but needs significant work to make it hang together effectively. The more cloud agnostic you make your solution, the more complex it will be - but in theory, the more benefit you will obtain.

As mentioned before, vendors like  ServiceNow, ServiceMesh, Cliqr and Chef will likely be involved if you work on this during 2015. Perhaps one of the major cloud vendors will introduce a service agnostic catalog offering in the future? Maybe it will come from Redmond. Stranger things have happened….Have a read here if you missed the Azure release earlier in the year permitting management of your estate across multiple providers. Can you imagine Microsoft offering that 5 years ago?

4) My fabulous colleague Jason Sutherland pointed out to me that once a self service solution was in place, the continuing Service Management would actually require significantly more design effort and ongoing work than the "broker" part of the model. As a hint of things to think about, you need to make the following work, at scale, across services, business units and cloud platforms:   

All very interesting, but I simply cannot live without a summary

No problem, I’ll help you out:

Self Service will be a major trend in cloud computing for the enterprise.

It’s not 100% polished yet, but you can position yourself in the right place to adopt this now. It won’t be easy, but the potential benefits are huge. Expect significant further developments in the coming years.