Private PaaS vs. Public PaaS - Bets Are Off

Hand Cloud

When looking at the (short) history of PaaS, one can split it into two main waves. The first wave (~2006) is best illustrated by Heroku, whose success was, in large part, based on the hyper-growth of two underlying phenomenons: GitHub and Ruby (and Rails). Starting from scratch, the best practices of this ecosystem have been defined in the cloud and are now core to its DNA.

The second wave of PaaS consists of vendors aiming at defining the more general problem of what next generation software development and deployment in the cloud and mobile era looks like. The distinction is important: this wave is about projecting existing developers, tools, processes, data and best practices in the cloud era, not starting from a blank slate.

If you look at the main middleware vendors (Red Hat, VMware, IBM and Oracle) from that second wave, all of them came out with a PaaS announcement in 2011. Those announcements, however, were not backed by true production-ready offerings.

Fast-forward two years and we are now seeing the delivery of v1.0 from some of those vendors. But what happened during those two years? One obvious part of the answer is that those vendors were simply busy implementing a PaaS solution to be more than a funky brand and a pre-recorded demo. But the less obvious part of the answer is that those vendors have, for a good part, spent a lot of IQ cycles wondering whether they should release a Private or Public PaaS solution. The debate is not as simple as it would seem, at first.

Current middleware vendors are selling top-down within IT organizations. That is where they excel. Steering their ship towards a completely different way of selling, measuring and compensating their sales force would create a huge disruption and risk loss of established revenue streams. Why take that risk when they have a chance to repackage their existing middleware and infrastructure offering as a combined packaged stack offering, the Holy Grail they have failed to achieve for the most part in the last decade?

On the other hand, those vendors do understand that the next-generation winning IT business model truly sits in the public cloud: this is where true CAPEX, OPEX and time-to-market gains happen. Those gains won’t happen for the legacy vendors by magically reshuffling existing hardware and software assets under a different brand and repackaging them. The power of the public cloud is harnessed by embracing the service era, where the tasks that don’t differentiate your company are externalized to cloud specialists that will do it at scale – and even better and cheaper – in a constantly evolving/always up-to-date environment. Changing titles on the hats in an IT organization won’t deliver that productivity boost.

One could argue that the other obvious option would be for those vendors to offer both a private and public PaaS. The problem is that maintaining both a PaaS software layer, that goes through 12- to 24- month release cycles (and that customers might decide to only upgrade every five years to match their IT cycles and budgets), while, at the same time, maintaining a public cloud offering – where the core value is to constantly evolve and improve – and keeping those two environments compatible is a match made in hell. Don’t be surprised if you see all of those vendors come up with both a public and private cloud offering. However, their public cloud offering will simply be demoware for their true area of focus: selling combined infrastructure and middleware stacks within traditional IT data centers.

If you think that what differentiates your company is not how good you are at patching Linux, installing load-balancers, maintaining drivers and application servers, and the speed at which you can cable servers, you should definitely Sacha Labourey CEO founder CloudBeesSacha Labourey is the CEO and founder of CloudBees. Prior to CloudBees, he was very involved in the JBoss community, first as a core contributor, ultimately becoming general manager for JBoss Europe and then CTO. After the Red Hat acquisition of JBoss, Sacha played a crucial role in integrating and productizing JBoss with Red Hat offerings. He became co-general manager for Red Hat’s middleware division before ultimately leaving Red Hat in 2009. He founded CloudBees in 2010. Follow Sacha on Twitter.