Cloud computing is growing fast. The leader, Amazon and Microsoft, continue to show more and more customers using their platforms every quarter. I do truly think that an increasing number of organizations are thinking cloud first for application development. That doesn't mean all systems, nor does it mean that they won't opt to bring their code back on premises, but they are considering the cloud.

Will this trend continue? For most of my life, I've seen trends come and go, but the popular ones continue to evolve and grow. I suspect that for now, cloud computing is attractive in many ways, not the least of which is a shift from a large up-front investment to a small pay as you go charge. Lots of finance people would rather see this as a way of investing in computing for their organization, knowing they can cancel and switch their investment elsewhere.

I think clearly the cloud will continue to grow for now, but will companies ever reverse the trend and move things back on premises? I wonder. It does seem that many people try to move to the cloud and struggle with the application changes, the unexpectedly high monthly costs, and the wandering performance levels of their systems. For executives used to knowing what the predictable level of performance was, even if poor, having systems (appear to) run randomly slower or faster can be maddening. I think some of this is the lack of understanding of cloud paradigms and programming practices for many developers as well as poor application architectures. But that's the state of the world, so if systems don't run well, why would you move away from your current data center?

The major cloud providers are providing lots of services and incentives, and certainly they are working hard to sell you computing resources. For some technologies, this makes perfect sense. I'd never set up an email server or system again, relying on a service like GMail or O365. I' not sure I'd even want to host a web server again, expecting that something like App Services is just out there, and I can drop my code in a location and I'll have a running website. And, of course, I love VSTS. If I were a developer starting a company, or a project, I'd consider using something like VSTS or a TeamCity suite with a cloud VCS. I mean, why invest and manage servers? I just want services.

There are issues with cloud computing, and certainly your flexibility can be limited. Security offerings are both better and worse, and pricing can be steep as you scale up. Of course, if you scale down you haven't spent a bunch of money on extra hardware. Overall, I think there are some sweet spots for cloud computing, and even though some workloads don't make sense, I'd expect that it's worth considering the cloud first.

I do expect that some services, such as email, may never come back on premises. I also think some of the skills and tools we use will change, evolve, or disappear. For everyone, I'd suggest you experiment and learn some things about the cloud. Get a free account to experiment with, download emulators, or just watch a lot of presentations and content on the cloud. Whether you have plans to move or feel pressure now, I'm sure you'll have to deal with the cloud at some point in the next ten years, so be prepared.

Join the debate, and respond to today's editorial on the forums

The median value is added to the plot with the abline function. The h lets the plot know this is a horizontal line, with the value equal to the median.

Ref: abline = click here

» Discuss this question and answer on the forums