I love the cloud. I’ve had the opportunity to run mission critical cloud services and applications at several large enterprises. I migrate workloads to the cloud. Sometimes, I migrate workloads out of the public cloud into a private one. In enterprise technology, we have an odd relationship with the cloud. We often get it right, but sometimes, we get it wrong. I’m writing this blog to level-set the discussion.

The cloud isn’t a Visio shape

Once upon a time the cloud was a Visio shape. Systems engineers and architects used to place it somewhere on their complex distributed application diagrams. We knew in great-detail what happened with each interconnected object on the diagram. We knew what went to the cloud and what came from the cloud, but we didn’t know what happened in the cloud. In general, the cloud represented “magic happens here.” It was a black box in which someone else knew the details. We were obscured from it.

The cloud isn’t a building

We often use the term “cloud” as the opposite of “on-premises.” Some software you can buy in either flavor. Usually the cloud version is better and the on-premises version is inferior. At least, that’s how they are marketed. This comes down to little more than geography. Is the software running on a computer where I can go and touch it, or not? Why does that question even matter? It’s still running on someone’s server. That server needs to be physically and logically secured. It needs to be patched, maintained, managed, and lifecycled. You can pay your staff to do this, or pay a provider to do it. Fundamentally, this isn’t a significant difference.

The cloud isn’t a financial model

The cloud gives you the special privilege of spending your dollars from the expense category instead of the capital category. This is often marketed as a significant advantage. Perhaps in small companies, this is a big deal. At scale, you can be flexible either way. Would you avoid spending $1.00 of capital by spending $1.50 of expense? Probably not.

In my experience, the cloud hasn’t significantly impacted financial flexibility positively or negatively. Yes, in theory, you can hyperscale a 1000-node mega cluster to compute for 1 hour, then destroy it, spending very little to do it. The reality is that the vast majority of enterprise workloads just aren’t that elastic.

The cloud isn’t someone else’s job

Historically enterprise technologists get a little too focused on what is right-in front of them. We take care of the technology in our data centers, and we pay someone else to take care of the technology in the cloud. The thing is, the cloud doesn’t run itself. It’s also not the turn-key black box that the marketing brochure said it would be.

In fact, modern cloud technologies are more open and tunable than ever before. The demand for on-staff engineering of cloud technologies is bigger than ever and will continue to grow. Cloud doesn’t make the business less dependent on corporate IT. If you are doing it right, the cloud requires more precise integration, data, and security engineering than anything before.

One person’s cloud is another person’s on-prem

This is where is gets really confusing. 10 years ago, I was hired to build a SaaS cloud for a major enterprise application. The application was developed in-house and distributed via CDs to our customers to install on their servers and workstations. I was part of a team that transitioned the application to an online multi-tenant, high-performance, elastic, and highly available system. For a while, we offered both, but eventually we deprecated the on-premises edition, stopped mailing CDs, and only offered our cloud SaaS model.

We ran the online application in our on-premises data center, on physical servers that you could walk up to and touch. Is this the cloud or is it not? Well, that entirely depends on your perspective.

OK, then, what is the cloud?

Thank you for your patience. I’ve spent this entire blog telling you what the cloud is not. So, what is it? Cloud is in the eye of the beholder, the end user. In a word, it’s an experience. I love to refer back to the NIST SP 800-145 Definition of Cloud Computing. From the view of the end user, your technology needs to have on-demand self-service, broad network access, resource pooling, rapid elasticity, and measured service.

What does any of that have to do with buildings, finances, and labor sourcing? Nothing. Actually some of this goes back to my original point. The cloud is a Visio shape, but from the end-user’s standpoint, not from the technologist’s understanding. The end-user wants it easy, fast, flexible, and measured. We are the ones that engineer it end-to-end, regardless of what building the technology runs in, how it was paid for, or what partners are engaged in the delivery. We own it. We are the cloud. Act accordingly.

Like my article? Want to read other articles I’ve written on cloud computing?  Check them out here:

Leadership as a Service: Five Ways to Ascend your Leadership into the Stratosphere

The Future of Infrastructure Technology