Platform as a Service is the least defined and smallest (annual spend) layer in the “* as a Service stack” despite being a growing market, but this is changing. Venturebeat explain the cloud layers well if you need a refresher – http://venturebeat.com/2011/11/14/cloud-iaas-paas-saas/.
Graph data from: http://www.idc.com/getdoc.jsp?containerId=prUS24298013
IaaS and SaaS dominate the market and are well understood. Use Google Apps or Salesforce? Then you’re using SaaS. Bought a virtual machine from a ‘cloud’ supplier? You’re using IaaS! The big gap is PaaS.
Now we’re seeing the same repeatable pattern of growth and development from the IaaS world in PaaS. Open-source projects are developing existing technologies (see Docker and LXC), then major vendors absorb / package / develop those technologies into (sometimes open source) off the shelf stacks. The big players are RedHat with OpenShift, and Pivotal (spin out from EMC and VMware) with CloudFoundry.
Think back to a maturing IaaS market before we began the race to the bottom we’re seeing now. If you wanted IaaS or API driven infrastructure then you’re buying public cloud.
Rackspace’s collaboration with NASA resulted in OpenStack, in my view the best private cloud option in the marketplace. It’s made up of multiple components and has become the defacto IaaS solution. Right now Enterprise are seeing it as a cheaper way to run a data centre. Not popular with open source advocates, but an OpenStack implementation can offer massive cost savings when compared with proprietary options – this will change no doubt. OpenStack is, funnily enough, open; there’s no proprietary lock in. It’s flexible, being built on interconnected components (compute, storage etc), and it’s API driven.
This maps neatly onto the developing PaaS world, but we’re yet to see a winner (OpenStack has won the IaaS crown, despite what its critics say!). So why is PaaS going to have such a big impact?
Software rules the roost whether it’s SaaS, backend applications, or websites and apps. IaaS (whether private or public) is a great solution for building software on – orchestrating infrastructure within an application puts controls in the developer’s hands. The problem is that there is still a lot of complexity here: networking, storage, backups, automation, deployment, high-availability to name a few. It takes pretty specialist knowledge to build an application solution on IaaS, and the time that developers spend building (and learning how to build) this solution is time not spent on doing what they do best: writing software.
PaaS allows them to do exactly that. By providing developers with a self contained, pre-configured application environment the complexity of infrastructure management is removed. In a PaaS world IaaS looks increasingly less relevant, unless you are a Service Provider – hello CatN! I believe that that software is a fundamental part of a successful business, PaaS will become a fundamental part of their IT strategy.
So what’s the end game? Well there isn’t one! PaaS continues to evolve (see xPaaS – the ‘x’ is variable), and the lines between middleware, integration systems, and application platforms continue to be blurred. As delivery of these technologies change, the way that businesses consume them will change too.
The key take away here is to remember why we build build websites and other applications. It’s about engaging with customers to delight them and hopefully retain them. Successful businesses are great at this, and it’s the same reason we’ve seen the rise of Agile an DevOps working methods. Continuous integration, iterative improvements / releases, and digital by default are all at home on PaaS.