Over the last few months we’ve become fascinated by the Platform as a Service (PaaS) market because after an initial phase where it was dominated by platform-specific and language-specific offerings, a set of Universal PaaS are emerging, many of which are based around the Open Source Cloud Foundry from VMware. In addition to PaaS, there is a class of vendors who provide external services to PaaS through “marketplaces” that the vendor sets up. We refer to these generically as Application Services as a Service (ASaaS). The stakes are potentially huge – the PaaS takes over from the Operating System as the dominant factor in the purchasing decision for server-side technology.  We’re not saying it definitely will happen, but it might.

Good examples of ASaaS are New Relic, which provides Monitoring as a Service, and Xeround, which offers Database as a Service. If you subscribe to any public PaaS there is a good chance you will be offered public ASaaS from New Relic  and Xeround through the vendor marketplace.

Quite often these ASaaS are offered as upgrades to a base-line service offered by the vendor.  For example, the PaaS vendor will typically offer you a bundled MySQL as part of your PaaS subscription, and a more scalable and available MySQL-compatible database from Xeround via the PaaS vendor’s marketplace. Similarly there will be some basic monitoring capability bundled with the platform by the PaaS vendor, but if you need better insights you buy a New Relic subscription through the PaaS vendor.

There is another interesting phenomenon – most of the ASaaS vendors offer a “freemium” model where a base version of their service is provided free.  They can then offer this back to the PaaS vendor  to replace the bundled equivalent of that service.  That would mean for example that the PaaS vendor’s bundled MySQL becomes Xeround Free under the covers (although the PaaS vendor doesn’t tell you this), and if you want additional function you can buy Xeround Basic or Xeround Pro through the marketplace. The ASaaS is being provided in both “white label” and branded forms to the same customer base. The benefit to the PaaS vendor is that it reduces the cost of developing and supporting the bundled services in the PaaS.  To give an example from the automotive industry, you may be able to pay extra to Ford to get Recaro seats factory-fitted to your new car, but in reality the standard Ford branded seats may also be made by Recaro in the same factory, to a slightly different specification.

At the moment there is incredible diversity and dynamism in the PaaS space, and it’s very hard to call the winner. It is, in a technical sense, a chaotic system in that small perturbations – e.g. some venture capitalist having lunch with some executive from IBM or HP may cause a company to be acquired which could change the whole dynamics of the space. It’s like the proverbial butterfly flapping its wings on the other side of the world, only with cutlery.

However, in a chaotic system there are usually islands of predictability.  It’s chaotic, not random, and it’s worth looking at things that aren’t sensitive to small changes in the dynamics of the system.  At this stage we think that the outcome in the  Application Services as a Service (ASaaS) market is much more predictable than the overall PaaS winner.  The logic is as follows

  • Certain ASaaS vendors have managed to establish themselves as the preferred ASaaS of their type in the marketplaces of a large number of PaaS players
  • Success breeds success – customers start demanding specific ASaaS vendors’ solutions from those PaaS vendors which don’t offer them
  • Whichever dominant PaaS vendors emerge, the same ASaaS vendors are dominant
  • At this stage this makes it much easier to “call” the successful ASaaS vendors – they’re the ones that have got the most traction  with the PaaS vendors.

If you combine the emerging dominance of certain ASaaS vendors in the PaaS space and the trend for white-labeling, you end up with something really very interesting.  Again by analogy with the automotive industry there are a very small number of suppliers of the white-labelled high-value components in a car. Whichever car you buy, for example, it is likely to have an ABS system from TRW, Bosch, or Continental.  These component manufacturers (although of very low public profile) have enormous revenues and enormous power, to the extent that they drive much of the innovation in the industry, and control much of the value chain.

So, if you are investing money in the space at the moment it is likely much easier to predict winners amongst the ASaaS vendors than the PaaS vendors and they end up with a near-monopoly and control of the value chain. I know where my money would be going (if I had any).