Cloud

Attack of the Cloud Killers:
Fending Off Threats to Cloud Ops Success

Print Friendly, PDF & Email

When businesses first started considering public cloud as a serious platform for IT, nobody was thinking too hard about how it would affect the organizational structure of IT shops or what the operating model for cloud should look like. But as cloud engagements have matured, those questions are coming forward in a big way, and IT organizations that haven’t yet put some serious thought into them run the risk of seeing their cloud projects shunted aside or derailed.

It was clear early on that many of the classic IT ops that central teams run – service ticket management, infrastructure health scanning, dedicated security operations, for example – would be going away. Many of those day-to-day functions are now the responsibility of the big cloud providers. That in itself was a substantial paradigm shift, and a difficult one to embrace for lots of IT pros who had lived and breathed those processes for many years of their careers.

Plus, the operating model that emerged – Cloud Ops – is very, very different. There’s a lot more automation involved, a lot more self-healing software that detects problems and fixes them. There’s a stronger emphasis on agility, speed and cost-effectiveness than what’s typical in the classic on-prem model. I know a financial services company that runs five thousand virtual machines in the cloud – a pretty decent-sized datacenter in anybody’s book. They run it 24×7, 365 days a year, with a Cloud Ops team of just 18 people, six to a shift. No security teams, no maintenance people, no-one walking the halls making sure the air conditioner’s running at the right level. If a server is not performing well, they just kill it and start up another one. They have exceptionally low overhead, while maintaining a very high level of service.

For more and more companies, the cloud model is no longer an option – it’s a mandate. If your competitor has an IT operation like the one I just described and you’re running a data center with 50 or 100 people, how are you going to compete?

Still, the transition to Cloud Ops can be challenging, no question. You’re up against all sorts of governance and insight issues and change management demands. These potential cloud killers are not so much technology challenges as organizational and cultural ones, and they all stem from the fact that Cloud Ops is a wholly different kind of IT world.

 

Here are three principles that I’ve found useful while working with companies on the front lines of cloud rollouts:

1. Understand the need for a completely different governance model.

What I have in mind here is not your classic governance, risk and compliance (GRC) processes – though those are crucial for cloud success too. Cloud Ops governance is all about how do I maximize value? It’s about having deep visibility into the financial health of the assets.

Finance controls are the number one thing that needs to be done differently from the beginning. As my colleague John Treadway pointed out in a recent post (3 Ways HPE GreenLake Hybrid Cloud Drives Hybrid IT Success), “if you’re not paying attention to what you’re using in a public cloud, you can easily end up overpaying.” What’s more, the acceleration of consumption is much, much faster with cloud than with the classic model, with all its POs, contracts, legal involvement and so on. Within a very short time, you can end up with uncontrolled, unmonitored usage and zero visibility. You’ve got spend that you can’t answer for, so what likely happens is that Finance steps in and kills the project.

Then you have a wrecked cloud initiative, and IT blames the business or falls back and says, “our cloud program cost too much” without actually looking for the root cause of the problem. It all could have been avoided by putting the right controls in place from the start.

2. Recognize that your Cloud Ops team can’t be the same team that’s running your classic model.

These groups need to be separate and differently dedicated. I may be taking a bit of a contrarian position here, because the IT leaders I talk to sometimes have a hard time getting their heads around this. And the decision as to how to populate each team is a tough one, for sure. But I’ve worked with quite a few companies that originally tried to merge their on-prem operational teams and their cloud operational teams in the hope that they could act as one. It hasn’t worked well at all.

The on-prem operational teams may not understand how the new model works. They’re probably not familiar with the technology and the new software platforms that you’re deploying. The cloud folks may lack the depth of experience needed to manage the on-prem assets. This is where the friction starts, with team dynamics issues, turf battles, silo building. I’m a convinced advocate for bypassing these problems by the simple expedient of keeping the two teams’ workflows distinct and separate.

3. Target your training.

A corollary to my previous point: it’s crucial to ensure that your cloud team has a clear understanding of the new paradigm. Maybe under the old model your staffers devoted a lot of time to the care and feeding of servers, ensuring their health and avoiding any need to turn them off. But now perhaps your new model is – and again I’m thinking of that financial services firm I mentioned – if a box isn’t working, get rid of it and bring in another one.

I’m not suggesting, of course, that your on-prem ops team won’t need access to training resources to continue learning and building their careers. But here again the two-team approach is useful. Cloud Ops training will be very different; it should be oriented more towards managing spend controls, enabling or disabling services, and supporting users who are consuming cloud services.

Importantly, it should also focus on the DevOps relationship and the benefits it delivers to IT service consumers. A tight connection between development teams and operations teams is pivotal for maximizing the value of cloud implementations. That partnership needs to be carefully fostered, and a big part of that is through training.

Building a top-flight Cloud Ops function is a demanding task, but an essential one for companies that want to see the best results from this innovative, agile paradigm.

 

Get Started Today

HPE has experienced consultants to support you through every stage of the cloud lifecycle. Get the expert assistance you need to quickly and effectively bring cloud computing services to your business. Take a look at our HPE Cloud Services website and get started today.

How useful was this post?

Click on a star to rate it!

Average rating / 5. Vote count:

No votes so far! Be the first to rate this post.

Leave a Comment