Tuesday, December 21, 2010

Now with database.com and heroku - how does salesforce.com stack up in the PaaS market?

While collaboration was a clear theme at Dreamforce with the launch of Chatter Free, Dreamforce ‘10 will be remembered for the serious saleforce.com play into PaaS. Over the course of just two Dreamforce keynotes, the PaaS market altered dramatically.  Salesforce.com launched database.com a database-as-a-service offering and announced the acquisition of Heroku a small but well-known Ruby on Rails cloud service. Along with some enhancements and re-organisation of offerings under the Force.com label, salesforce.com invested their brand fully into the PaaS market.

While SaaS serves the end user, PaaS serves the developer community enabling rapid solution delivery on an elastic infrastructure. However, the audience for PaaS is somewhat wider than the traditional development community. In addition to the usual Java and .NET code junkies, some PaaS offerings such Workxpress and Orangescape strive to make development more simple through “drag and drop” style configuration. Such configuration tools will appeal to the cottage industry of developers that work "on the business side". Just as MS Access and Excel enabled Andy in front-office to build that interim project app which ICT have spent the last years trying to kill off, the cloud gives more powerful and easier to use options (that ICT could spend many more years trying to bring under control!).

So having set the standard for software-as-a-service, how does salesforce.com now stack up against the competition in the PaaS market? It’s a perfect time to take stock of the present landscape and what it offers.

I’ve done a bit of research and compiled these offers into a high-level table of what I see below (I’ve chosen just 6 of around 60 offerings in the marketplace. This is based upon the visibility of the first three App Engine, Azure and Force.com and Workxpress/Orangescape have attracted analyst attention and are interesting propositions):

  • Most of the data in the above table has been sourced from publicly available information although not so easy to track down and I hope that bringing this together is useful
  • I have kept Heroku separate as it will continue to be supported as a separate cloud from the Force.com platform at least for the near-term
  • Adoption is difficult to measure. Most PaaS providers are indicating no. of applications but obviously for those that are looking for an indication of readiness for enterprise adoption (like me) the size and complexity of those applications is a measure. Same goes for number of customers.
From my latest look across the PaaS landscape I drew a few conclusions:

1.   PaaS offerings vary widely and do not compare directly – therefore there is not a  “best platform”
When looking across the PaaS providers the first thing that becomes clear is that they all have different offerings. This should be expected in an immature market where organisations are still interpreting the client need and trying to find a unique proposition. The most traditional services from a development approach perspective – App Engine, Azure and Heroku support straightforward coding in 3GL programming languages. App Engine and Heroku are more specifically focused on the highly-scaling web application (and have constraints to match) while Microsoft has the broadest offering and is trying to appeal to the its present day development community to adopt its cloud platform. Outside of the 3GL coding space there are also 4GL tools available – Force.com extends the popular salesforce.com app, Workxpress is a drag-and-drop app builder on its own native environment and OrangeScape is a combination of drag-and-drop with scripting and deployed to a user selected public cloud. It’s important when assessing the PaaS space that the types of applications you want to build will heavily influence the best choice of platform for you (even if you are deciding between the 3GL providers). See this very interesting link to appreciate how the wrong choice (in this case Google App Engine) can kill a project. While obvious to the tech-savvy cloud computing fans, I have seen many bloggers write along the lines of “developers won’t wait for VMForce Java on Force.com they will go straight to Heroku” as if there is a straight-forward comparison between the 3GL platforms.

2.   The most interesting PaaS use cases are coming out the Microsoft camp Specifically Azure RenderMan (the comments are also amusing!) and Transport for London

3.   Many PaaS offer ‘for-free’ models to attract a development community
Most cloud vendors in this space offer at least a free trial phase or free limited volumes in order to encourage developers to test out the capabilities. This is key for success particularly as grass roots developers adoption is the name of the game for getting a standard in place. A larger development community on a platform really increases the attractiveness of a platform as an eco-system of ISV applications are made available and development patterns / knowledge is spread. For that reason, IT execs and project budget holders will not be the ones that decide that they will go Heroku or they will go Google App Engine – they will look to the in-house developers experience with these dabbling in these platforms or out to the market on what is easiest to resource. It will be the developers who ultimately “choose” the platform they prefer. This is much different to the world of Enterprise Applications where SaaS plays in which deals are made with big brands through senior relationships.

So is the new offering from salesforce.com well positioned against this competitive landscape?

Yes I think so.

Whether intentional or not, the services under the new Force.com umbrella (Heroku, Database.com, VMForce and the original Force.com platform) stretch across the spectrum of potential PaaS tools. These tools will work as standalone options serving specific use cases in the PaaS market. Heroku will cater for rapid development of highly scalable web apps, and the forthcoming VMForce (java) arrival will cater for broader application development needs. I don’t see an obvious integrated use case that connects all of the salesforce.com PaaS offerings but in my view that is not an issue. By providing everything from SaaS to low-level PaaS, salesforce.com has a cloud solution for many different point problems and may become the best of brand choice for an enterprise looking to utilize the cloud to it’s full potential. Given that with each additional cloud provider used comes an even greater multitude of additional risks (data, security and infrastructure ops will go offsite), businesses will want to minimize the number of cloud service providers even more than most type of suppliers. For this reason, salesforce.com’s dramatic increase in breadth of offering combined by its reliable brands (Heroku is a well reputed service) positions them very well for the future.

As 2010 draws to a close, I’m expecting the new year to bring much more activity in the PaaS market as the major players begin to move in. Looking forward to it!

No comments:

Post a Comment