5 cloud deployment blunders to avoid

I’ve been consulting on cloud migrations and deployments for a while now. Along the way, I’ve seen lots of mistakes made. Mea culpa: I’ve made some of them myself.

Don’t be like me, or a lot of companies. Avoid these five blunders and you’ll be much better off.

1 Hurry up and fail.

Do not, I repeat, do not, run into the cloud as fast as you can.

You’ll only fall and hurt yourself.

Sure, a recent Uptime Institute survey of 1,000 IT executives, revealed that 50% of senior enterprise IT executives expect most of their IT workloads to reside on the cloud by the end of the decade. That doesn’t mean you need to rush into it.

If you charge headlong  into the cloud without considering all the issues — from costs to regulatory issues to workflow–you will create more problems than you solved.

Take your time, do your homework, run extensive test projects — and then slowly and deliberately start your move.

2 Count your cash savings before they’re hatched.

Yes, you can save money by moving from data centers to the cloud. But, there’s nothing magical about moving from your old IT Capital Expense (CapEx) spending to Operational Expenses (OpEx).

It’s apples and oranges. Until you’ve done a complete financial analysis of what you’re spending now and what you’ll be spending on a cloud model, don’t jump into a long-term cloud contract. What might look like a great deal at first glance may turn into a financial dark hole.

Keep in mind that what you’re really buying with a cloud is flexibility. You can often save cash, but unless you’re careful, you can still blow your IT budget.

3 Think all clouds are created equal.

Amazon Web Services (AWS), Azure, Google Cloud, Oracle Cloud, IBM Cloud Bluemix, the public cloud list goes on and on. Besides offering cloud services, they have one thing in common: They don’t offer the same services, service level agreements (SLA)s or even the same pricing structures.

Clouds are like apples. They come in Honeycrisp, Red Delicious, Granny Smiths, WineSaps. They may all look alike, but after one bite, you’ll know there’s an enormous difference between them.

Then, when you throw in private and hybrid cloud, things become even more complicated.

What kind of cloud do you want? Everyone offers Infrastructure as a Service (IaaS), but what if you want a cloud for your electronic healthcare records? The big guys probably can’t help you, and you’ll need to look at smaller, vertical cloud providers.

In short, don’t just jump on a cloud because the initial request for query (RFQ) looks great or it’s a vendor you used for years. Analyze exactly what the cloud provider or private/hybrid cloud partner can give you and how that fits in with your needs.

4 Don’t consider your connection speed.

Software as a Service (SaaS) is a great cloud model. But, before you move all your legacy desktop applications to SaaS, ask just how fast is your Internet connection?

Can your corporate headquarters net pipeline handle thousands of workers creating documents or spreadsheets on Google Docs or Office 365 simultaneously? Do you have off-site workers? Are some of them still stuck with 5 Mbps DSL connections? Can they do their work with Salesforce CRM?

If your Internet connections are slow, a SaaS is not for you. It may also mean that, for example, Disaster Recovery as a Service (DRaaS) or backups may not be for you either.

A cloud is only as good as your Internet connection.

5 Give up control of your own data.

Cloud vendors are only responsible for uptime, and most SLAs won’t even guarantee that.

As Lydia Leong, a research vice president in the Technology and Service Providers group at Gartner, recently observed, “Cloud IaaS SLAs can readily be structured to make it unlikely that you’ll ever see a penny of money back.”

So, make darn sure your data is backed up on another site or locally. That way at least you’ll have your data.

You’ll also want a robust disaster-recovery strategy. For example, if you can afford it, you want your programs running on two availability zones if you go the public cloud route.

You may have noticed a common theme with all these problems. I’ll spell it out for you: Take Your Time.

Moving to any cloud, whether it’s a major public cloud or making your server room a small private cloud, is a major decision. As it is, according to an IDC survey 40% of businesses with public cloud experience moved back to on-premise computing. You don’t want to explain to your CFO or your board why you turned out to be one of them.

Treat the process with due diligence and you’ll be much happier in the end.


The 5 biggest cloud developments to watch in 2017

Who should be making your cloud decisions?

Hints and tips for developing your cloud strategy in 2017


  1. Nice one! Thanks for this, a good summary.

  2. Cloud applications are not the same as on-premises applications.

    If you try to lift and shift, you’ll run into the issues noted as well as re-inventing the wheel (probably designing a stagecoach wheel when you need something for a car). Follow the capex timelines and migrate an application at a time: the finances don’t work otherwise as there’s still sunk cost in the old capex investments.

    You’ve got to build new IT capabilities, and kill some old ones. Both changes take time and care.

Speak Your Mind


This site uses Akismet to reduce spam. Learn how your comment data is processed.