My excellent cloud adventure: buggy whips and other hard CIO challenges

May 31, 2017   |  

There’s an old urban legend that says that when the first automobiles came out, they included whip holders. The idea was that people had always used whips to drive horses on their old buggies, so they’d probably want them for their new cars as well. Yes, it sounds ridiculous. But did you ever stop to wonder why Tesla puts radiator grills on their electric cars?

Paradigm shifts create some funny behaviors. And this is only logical, as people look to capitalize on a promising yet unexplored new opportunity while keeping one foot grounded in the present.
No where do I see that truth reflected more than in Microsoft’s transition to move our entire IT footprint to the cloud.

As I wrote in an earlier blog, it can make otherwise brilliant engineers come off looking like sloppy teenagers. Today, I want to talk about how it changes the life of the CIO. It’s not a stretch to say he or she needs to question – and be prepared to blow up – almost every aspect of IT they ever believed to be true.

Take, for example, the basic standards by which CIOs have represented their portfolios to their boards and executives for the past 40 years. In a broad sense, they’ve fallen neatly into general, established categories: app count, server count, and cost (comprising both “CapEx,” or Capital Expense, and “OpEx,” or Operating Expense, which includes people). The smart CIO understands that the cloud makes these kinds of distinctions obsolete. In fact, holding onto them can leave you stuck hopelessly in the past.

Consider the very concept of an application. It’s easy to grok that moving to IaaS changes the underlying infrastructure. It gets a little fuzzier to think about how the move to PaaS changes things. But in true cloud world, the fact is that the concept of “app” effectively melts away.

Let me give just one example, based on our experience in Microsoft IT. For years, we provided an app called Paystub. It’s where people went to find out exactly how much money they made, how much they paid in taxes, and how much vacation time they had. In the old days, we stood this up with six on-premise virtual machines, two databases, and a web front-end. Every two weeks, when the paychecks would go out, SAP produced a data dump that would go into a distribution system we called “feed store,” and the Payroll app would copy it into the two databases, from which it was presented to employees when they used the app.

This worked reasonably well. We had infrastructure and an application – and we also had one thing we didn’t necessarily want: multiple copies of very sensitive personal information that needed to be secured and protected.

This all changed when we embraced the idea of moving to the cloud. We wanted to do much more than just get rid of infrastructure. And it turned out that we had an interesting opportunity in something we were building elsewhere called HR Profile. This was a PaaS-based solution designed to give people a single place to find and update all their personal employee information. We said: That’s the perfect place to put Paystub, right? So we did. And now, when you go there and click “Paystub,” there is no database behind it. No servers. No VMs. Instead it makes one call all the way back to SAP to get your information and shows it to you when you want it. And when you close that web page, it shreds the data and throws it away.

You could argue that there’s still Paystub functionality – that it’s now just part of something called HR Profile, but is it still an app?  I’d suggest it’s better to view this as an experience. There’s no server count, no infrastructure. It’s just a PaaS subscription with a cost behind it. And we’re constantly adding new aspects to that experience.

How should the CIO think about this? What does he or she report? What’s in his or her portfolio? You could say it’s a collection of experiences, but even that’s tough because it feels arbitrary and quite possibly limiting. Who’s to say where the boundaries of a given experience should end? And as you stand up more and more of your core of IT as services in PaaS with exposed APIs, you can end up with any number of unique experiences that composite those services in unique ways – making counting/describing the portfolio even more complicated.

So, are there apps anymore? Almost all the IT toolsets we’ve used for generations start with “What’s the app name?” and then “What’s the associated infra?” This all gets destroyed. Is the CIO’s portfolio now just an Azure bill and subscription count?

These are questions that every smart CIO and IT leader is asking today. In the new world, we must find new ways to represent value and strategic impact to the organization. We’re all dealing with it.
Referring back to our car analogy, it’s time to start imagining not how to build a better automobile but rather a world in which the idea of “car” transforms, and we think not in terms of the vehicle we own and maintain but rather the transit service we purchase. And so it is with IT.

Nobody wants to be remembered as the old leader who installed buggy whips on the new horseless carriage. Don’t let it be you.

Tags: ,