3 PaaS lessons from CloudBees’ funding

I’ve been following Java-centric Platform-as-a-Service startup CloudBees for a while now, so I wasn’t too surprised when I learned it had closed a $10.5 million Series B funding round. PaaS is a white-hot space in terms of developer and VC interest, but it’s also very competitive. Thus far, however, CloudBees has demonstrated an impressive strategy for differentiation.

Other PaaS startups might want to pay attention:

Be specialized. And I don’t mean just support a single language, I mean support a single language and have some serious talent to lead the development charge. As I noted when covering CloudBees’ $4 million Series A round late last year, it includes many former JBoss executives and developers among its team, and JBoss founder Marc Fleury is among its investors. If you’re going to do Java, there are worse guys to have on your side.

We’ve seen this tactic deployed by numerous other PaaS startups, too. Heroku (s crm) recently brought on Ruby creator Yukihiro Matsumoto to lead development of its flagship language, and Joyent now employs Node.js creator Ryan Dahl.

The future of PaaS might be multi-language/framework, as Heroku’s recent support for Node.js and Clojure, as well as VMware (s vmw) Cloud Foundry and DotCloud, illustrate, but single-language will still work. It just has to be done well.

Be inclusive. I mean this in terms of both products and potential user base. CloudBees supports Java developers throughout the application life cycle with two distinct offerings: [email protected], serves up the Jenkins (nee Hudson) development platform as a service, and [email protected], which is CloudBees’ production application runtime environment.

Furthermore, CloudBees has expanded the number of infrastructure options its service can run on to include OpenStack and VMware vSphere in addition to the original Amazon Web Services support. This is a big deal because whereas many PaaS offerings are relegated to running atop a single IaaS cloud, CloudBees users can choose between different public clouds, or even can run the service as software in-house on private cloud infrastructure.

Be first. Here’s the other thing about CloudBees: It wasn’t the first PaaS provider to focus on Java, nor was it the first to offer its service as software that can run on multiple environments. But, as far as I know, it was to first to market with both capabilities.

Already in talks to merge with Java PaaS pioneer Stax Networks, CloudBees sped up the culmination of that deal in late-November after Red Hat bought the JBoss-based Makara. Stax’s service was already running, so CloudBees was able to offer a product earlier than initially expected, while working in the background to integrate its technology. In June, it rolled out a Premium version of [email protected] to complement its free version enhanced features.

All of this happened while other Java-based platforms such as VMforce still look like vaporware and Google (s goog) is still tweaking its enterprise-friendly App Engine with full Java support. Another startup, CumuLogic, is doing PaaS software that can run atop a variety of infrastructure options, but it’s still in beta. Red Hat (s rht), too, took its time to productize the Makara platform as OpenShift, which was only recently released in Developer Preview phase.

Of course, it also helps to be a quality platform. CloudBees recently fared well against Google App Engine and Amazon Elastic Beanstalk in a third-party review of Java PaaS offerings.

None of this is to say that CloudBees is or will be the premier Java PaaS provider. Just that it’s not surprising to see it attracting investors’ money.

Image courtesy of Flickr user yomanimus