12_tribes - Fotolia
Independent PaaS providers, such as CloudBees and DotCloud, have shifted their focus -- the former to Jenkins and the latter rebranding as Docker. Others, like CumuLogic have disappeared altogether. More commonly, these PaaS vendors have become part of a larger cloud computing layer, as was the case for HPE acquisition Stackato, as well as Heroku, which is now owned by Salesforce.
Meanwhile, the public IaaS vendors have deemphasized many of their early PaaS tools amid flagging adoption. Some industry observers attribute this to the products being too early for the market, while others argue the platforms were too closed and proprietary. Regardless, PaaS has never truly lived up to its early billing as the future of cloud computing.
Today, PaaS remains a pattern of cloud computing -- public or private. On that point, there is no argument. What's less clear is where the market is headed and whether PaaS can continue to provide value.
A brief PaaS history lesson
To understand what led us here, let's go back to 2008 for a moment. The National Institute of Standards and Technology was forming the way we defined the types of cloud. IaaS was for workloads traditionally found in the data center, such as storage and compute platforms; SaaS was for static business processes presented as over-the-internet applications; and between the two there was PaaS.
PaaS, at its essence, was and is a development and deployment platform. Organizations use PaaS to build applications with standard programming languages -- Node, Python, Java, etc. -- linked to the provider's database, or to a third-party database, then deploy the applications for production. Providers accomplished this in different ways, but PaaS clouds were and are largely closed systems that many describe as a development "sandbox."
Some PaaS players were more enterprise focused and open source focused. For instance, Red Hat OpenShift is an open source container-based platform focused on the private PaaS market. OpenShift provides developers with an integrated development environment for building and deploying Docker-formatted containers, with the target deployment platform being Kubernetes.
Another pure-play enterprise PaaS that emerged earlier this decade was Cloud Foundry, which is open source and supports Java, Node.js, Go, PHP, Python, Ruby, .NET Core and Staticfile. Cloud Foundry was originally created by VMware, and commercial efforts are currently spearheaded by companies such as Pivotal Software, a Dell Technologies company. Cloud Foundry has also integrated with Kubernetes.
It's also worth remembering that Microsoft and Google's initial forays into cloud were focused on PaaS but both quickly moved to IaaS and have kept their focus there since. Meanwhile, AWS, which began with IaaS, later added PaaS with Elastic Beanstalk.
A PaaS postmortem
There are multiple reasons the PaaS market didn't play out as some expected. For starters, most of the focus was on infrastructure -- not development -- when these platforms emerged. In retrospect, PaaS vendors were perhaps too aggressive in their attempts to provide approaches and technologies to innovate the cloud application development game.
More importantly, many of the advantages of PaaS -- simplicity, portability, lower overhead -- have been diminished by the emergence of serverless computing and Kubernetes, which offer similar capabilities, to varying degrees. Moreover, the lines between IaaS and PaaS have blurred to near invisibility as AWS, Microsoft and Google continue to add features and functionality that fill the gaps between the two cloud computing models, particularly around app development.
As a result, PaaS vendors that aren't a part of a larger IaaS platform have a diminished role in the cloud computing market. Most quietly faded away or pivoted to become something else. For example, the aforementioned CloudBees successfully refocused around DevOps. Perhaps executives there understood there were too many larger players in the PaaS market, and that they needed to focus more on a niche.
Where does PaaS go from here?
PaaS has become so layered into the larger set of offerings from the public cloud providers that it's rarely referred to anymore. These vendors offer IaaS, PaaS and SaaS, but expect to see those acronyms used less often as the clouds get bigger and their patterns increasingly conflate. In fact, it's gotten to the point where the term "PaaS" is itself considered toxic. Once proud PaaS providers seem to be moving away from it slowly, not even displaying the term on the homepages anymore.
I'm not sure any of this is surprising. But what does surprise me is the fact that the term "PaaS," and the companies that focused on PaaS, remained so long in a market that's changing faster than any we've seen in the history of IT. There's still room for the remaining small independent PaaS providers today, but at some point they may need to refocus their technology in other directions that generate more revenue, like CloudBees did years ago.
AWS, Google and Microsoft will continue to appear as juggernauts in the marketplace, displacing the need for independent PaaS. Yes, it's quickly fading as a term, and as a stand-alone concept. However, it made valuable contributions to today's modern public clouds. Let's focus on that.