There are lots of big cloud shows coming up, and the core themes will be containers, devops integration, and more serverless computing services, such as databases, middleware, and dev tools.
Why the focus on serverless computing? It’s a helpful concept, where you don’t have to think about the number of resources you need to attach to a public cloud service, such as storage and compute. You just use the service, and the back-end server instances are managed for you: “magically” provisioned, used, and deprovisioned.
The serverless cloud computing concept is now white-hot in the cloud computing world, and the cloud providers are looking to cash in. Who can blame them? At the same time, you can take things to a silly level. I suspect there’ll be a few serverless concepts that jump the shark the day they are announced.
So, what is good serverless? The answer: anything having to do with development. Tools such as AWS Lambda and Microsoft Azure Functions are leading the way. As a developer myself, and someone who’s managed developers, the less I need to do to size up the back-end resources needed to run an application, the better.
While some people argue that these tools remove fine-grained control of the infrastructure, and could actually cost you more, my experience is that the more you automate and abstract from the developers around infrastructure, the better and cheaper it will be. The productivity gains alone pay for itself, if you bother to gather those metrics. I have.
On the other end of the spectrum, what’s an odd or silly use of serverless computing? Anything that’s not directly related to development, such as security, ops, and governance. The likelihood is that we’ll see the term “serverless” used to deal with how that you manage the back-end resources for these tools. Although it will be an improvement, the concept of allocating resources to support those tools while running should be a feature improvement, and the cool-kids term “serverless” slapped on.
I view serverless as something that should have been built into the public cloud services from the get-go. It surprised me how much detailed infrastructure stuff you had to initially do just to get things working. So, adding serverless now seems more like a fix for a long-time problem than an enhancement.
It’s inevitable that we’ll see serverless cloud computin g taken to a silly level in the marketing that now follows cloud. Don’t get too caught up in something that’s important in some aspects, but not in most others.