Will PaaS Become the Delivery of “Old Middleware in a New Way?”

By Atos Apprenda Support

PaaS has made a big promise to enterprise IT: the promise that things will be different this time around. Use a PaaS software layer on top of your datacenter and public IaaS (for a hybrid scenario) and the vision of a self-service, non-fractured IT stack will become a reality. Enterprise IT will no longer have 10+ platform stacks (“the Websphere folks,” “the .NET folks,” etc.) with disjointed operations and management because PaaS will consolidate them into a single point of operation, smoothing over those fractures. This leads a customer to the ultimate value of driving down the costs of operating IT by ten-fold by breaking down silos and homogenizing the operating model.

What does it take to accomplish this?

CustomerhatThis is something we think about at the product and market strategy level on a regular basis. Part of good product and market strategy is having empathy for the customer’s current state and their ideal state. If we put our customer hat on, of course we’d want a PaaS that consolidates our past successes and minimizes the significance of past mistakes. In order to do that, a PaaS software layer must be a “Switzerland,” showing impartial support of downstack components. It needs to have a business and architectural model to slowly plug into all the old vendors and be the hub in the hub-and-spoke system that a good PaaS implementation creates. Companies like Apprenda can do this because we are not beholden to old products that are still generating 99% of our revenue. We are a new company building a new world, not an old company trying to find a place in a new world, while having an identity crisis in the process.

What are big vendors doing in PaaS? They’re using PaaS as a way to extend the life of their flagship investments. To understand this, let’s look at some examples:

  1. VMware & EMC spun-out and co-own Pivotal. Pivotal offers the Pivotal CF PaaS that has an open-core which foundationally leverages the open source version of Cloud Foundry. Although the open source project is independent of infrastructure, as far as I can tell Pivotal CF has a requirement that you run VMware. I’m deducing this by the fact that their installer seems to explicitly call out, and maybe even install, ESX. This feels like a play to kowtow to parent companies VMware and EMC and to provide a vehicle for moving strategy into a new world without leaving the old virtualization model behind.
  2. RedHat’s OpenShift product is finely tuned around JBoss and RHEL. It doesn’t have any explicit support that would be considered architecturally equal to any other application server or distro and it puts its eggs in the OpenStack basket on the IaaS side. This feels like a play to drive more RHEL and JBoss EAP.
  3. IBM recently announced BlueMix, a public PaaS using the Cloud Foundry project at its core for some basic PaaS DNA. IBM used that DNA to assemble a WebSphere and IBM middleware PaaS. I conclude from this that it’s likely they’ll eventually release a software product. This feels like a relevancy play to ensure WebSphere has a cloud story.

None of these PaaS offerings can be a true Switzerland. RedHat will always claim that you shouldn’t run JBoss apps on any other PaaS and IBM will do the same for WebSphere. In fact, they may not really support any other app server other than their own, making it impossible anyhow. Pivotal will tell you that Cloud Foundry is infrastructurally-independent, but will encourage use of VMware. And each of these companies will have sales & marketing organizations tuned to these legacy-minded outcomes.

If you play this out to its logical conclusion, each vendor will essentially make their PaaS the best way to work with their old middleware products. None of these vendors provide new middleware; they provide old middleware in a new way. This means that there is a potential future where customers will end up running one PaaS per vendor, resulting in a future that is as fragmented as the past. No customer wants that. Customers want freedom from complexity.

Legacy revenues that fuel these companies have a gravitational pull that causes new product strategy to veer away from customer need so as to avoid the painful cannibalization that would occur by properly dealing with the innovator’s dilemma. The avoidance of this dilemma is levied as an implicit tax on the customer who will be asked by each of these vendors to continue to operate in a fragmented way.

As Apprenda continues to evolve, we subscribe to the vision of PaaS as an opportunity to move away from this fragmented past where each vendor and project creates a silo. Being unencumbered by legacy revenue means we can make decisions about our product where the “customer is first” and create a scenario where our past determines our customers’ future.


Atos Apprenda Support