I’ve been asking myself a question lately: is Platform-as-a-Service a disruptive technology? I’m not talking about whether PaaS will change things or is really different – it’s obvious that it has the potential to make a big difference. I’m talking about disruptive in the sense that Clayton Christensen introduced – the definition on Wikipedia is “an innovation that creates a new market by applying a different set of values, which ultimately (and unexpectedly) overtakes an existing market”.
This is an important question. If PaaS is a ‘sustaining’ innovation then we can expect a fairly smooth transition from the use of Infrastructure-as-a-Service through to Platform-as-a-Service. Existing companies in the IaaS area will be incentivized to invest heavily in PaaS and will start making money right away, probably from the IT departments in large companies.
But if PaaS is disruptive, that means that it will start out not meeting the minimum requirements for the mainstream market, making sales to mainstream customers (read corporate IT departments) difficult. At Appsecute we’ve experienced ourselves the benefits of PaaS – the potential is there to vastly reduce the costs and time involved in operating production systems. But if PaaS is missing key capabilities then corporate IT simply can’t buy it, no matter what the potential cost savings and benefits in agility. They have to stick with IaaS, with its greater maturity of tools and greater amount of control over the technology stack.
Disruptive technologies first find a foothold in another, non-mainstream market that has a different set of criteria for success. Typically disruptive technologies look terrible when evaluated against mainstream criteria, right up until they don’t. So disruptive technologies are consigned to marginal or niche markets where they gradually increase their capabilities, until one day, suddenly, they do have enough capability to be used in mainstream markets, and still have their unique benefits. Their traditional technology competitors are typically far ahead in capability by this time, but suddenly the disruptive technology is good enough on the mainstream criteria, and offers some other compelling advantage (cost savings, or some unique advantage) and all of a sudden the mainstream market flips, ditches the existing technology and adopts the disruptive technology en masse.
So let’s look at PaaS. Right now it’s missing some things that IT really need for production applications. Compliance with industry regulation, security standards, audit standards, repeatable and rigorous deployment processes – most of these things are missing from PaaS right now. I would go so far as to say that PaaS isn’t usable for production corporate applications.
But PaaS is finding a home in niche markets. It’s being used for development and test environments in corporates. And it is being used by the devops movement to run production applications outside of the control of IT. In these environments it doesn’t matter that PaaS isn’t ready to meet mainstream IT requirements yet.
PaaS is increasing in capability every day. At Appsecute we deal closely with Cloud Foundry from VMware, and the pace of innovation in the Cloud Foundry community is astonishing. And one day it will have enough features to be used for mainstream production applications. At Appsecute we’re doing our bit to help it along, by adding some of those features via our management tools.
I don’t know if PaaS really is a disruptive innovation, but I’m seeing a few signs that it is. And if that’s true then we are likely to suddenly see a transition from PaaS as a tool for devops to PaaS as a mainstream environment.
If PaaS is disruptive then at some stage in the near future we will see explosive adoption, replacing the use of Infrastructure-as-a-Service. It won’t be linear, it will be exponential.
If PaaS is disruptive then it’s going to be an interesting future for the cloud.