Sometimes I think we’re overdoing it with serverless. Once it was a handy platform that saved you from having to size server resources correctly and that removed resources when no longer needed. Today it’s a catchall that has different meanings depending on its applications and who’s provided serverless as a service on the public clouds.
I’m seeing some obvious trends that enterprises may be able to exploit. Here’s a head up.
It’s a Kubernetes world, and serverless is now well represented. Kubernetes has become the de facto open source container cluster and orchestration platform. Its movement to serverless was a foregone conclusion. Features such as networking, service discovery, and scaling, along with cluster and container federation in support of multicloud, are the benefits we’re likely to see emerge this year and next—all serverless enabled.
Statefulness may be a bad word among many newer developers, but serverless does support statefulness. In the recent past this was thought to be oil and water, given that serverless systems execute a serverless function in a stateless, not stateful, way.
However, many applications, including those moving from traditional systems, leverage state. Although stateful applications typically could not be ported to emerging serverless platforms, the emerging direction appears to be changing. This will likely clear the way for many more applications to move to serverless than originally thought. Many of the serverless technology limitations we dealt with just a few years ago are no more.
Those of you already immersed in devops know that Gitops is a version-control system. The idea is that Gitops will improve operational workflows for more agile and speedy projects. The integration between Gitops, serverless Kubernetes, and even functions-as-a-service will provide benefits not yet understood, and these systems will be largely public cloud-based.
The integration of serverless approaches and databases is already underway but will expand this year and next. I never understood why I had to size storage and other resources needed for a database system when serverless would take care of that, behind the scenes. I suspect that most enterprises will only accept a serverless approach to databases within two years’ time.
Of course, many other serverless trends will emerge as well, as cloud technology providers attempt to make their technology buzzword compliant. However, beyond the “serverless washing” that we’ll see more of in the near future, the core value of this technology is well known, and the problems being solved are pragmatic.