Is Cloud Foundry the Next OpenStack?

Disclosure – I am an adviser to ActiveState, creator of the Stackato Cloud Foundry-based PaaS. I’ve also consulted to a number of companies mentioned in this article. Full details on my disclosure page.

Ever since Pivotal was created with IP from both EMC and VMware, the cloud watchers have been closely observing to see what it means for Cloud Foundry, the Open Source PaaS created by VMware and now part of the Pivotal grouping. For those of us who believe that PaaS is the future of the cloud, this is a pretty important time – while we may firmly believe that PaaS is the future, it’s fair to say that we’ve not seen massive uptake of PaaS as the new paradigm for enterprise IT. There are lots of possible reasons for this – PaaS is far more difficult to understand than IaaS (the metaphor of IaaS being just like physical servers but acquired on a utility basis is a pretty simple concept). Add to this the fact that, since PaaS abstracts more of the stack away from them (one of the core benefits of PaaS) their decisions around architecture need to be more solid, and the ongoing fears around vendor lock in and you have a few big barriers still getting in the way of PaaS adoption. As barb Darrow pointed out in a GigaOm post last week:

Developers often love PaaSes, which give them an easy environment to build their applications, but big companies often balk at deploying those applications on a third-party platform and bring them in house.

Which is why it’s been fascinating over the past few weeks to see Pivotal strongly accelerate the building of a broad industry ecosystem around Cloud Foundry. While there have been, since the launch of the product, a number of smaller players offering Cloud Foundry based PaaSes (Tier 3, AppFog, ActiveState), it’s only been recently that we’ve seen large vendors really get on board. First came the announcement that IBM and Pivotal were forging a strong alliance around Cloud Foundry – including a deep commitment to IBM contributing to the Cloud Foundry product stable.

Last week it was the turn of CenturyLink – already inside the fold via its acquisition/acquihire of AppFog. The company announced that it is joining the Cloud Foundry community advisory board. These sort of advisory boards can be seen as nothing more than over-hyped coffee circles but in an open source project such as Cloud Foundry they’re actually pretty Pivotal (pun intended). These sort of cross vendor committees are often the first step to a full independent foundation to run the open source project. It’s a great model – while the commercial entities can focus strongly on building out their revenue models and unique product offerings, the foundation can look solely at ensuring the robustness of the project itself, and that a level playing field exists for all members.

As for the large vendors, the comments I made about IBM hold true for many others – as I said previously (and feel free to swap out the IBM reference to one for Dell, or HP:

…IBM, alongside other traditionally hardware-centric companies like Dell and HP, derives most of its revenue from selling infrastructure. And there’s one massive direct threat, and a large indirect one, looming for that. Amazon Web Services is pulling a massive amount of the general application hosting market that IBM might once have been able to sell hardware to. Not only are they doing so with their public cloud, but the recent on-again/off-again AWS winning of the Federal Government private cloud contract was a blow that must have huge for IBM, their competitor in the tender process. If that wasn’t bad enough, the inexorable rise of AWS has shown a generation of IT decision makers that building infrastructure on top of commodity hardware is a safe tactic – the result of this has been awesome for the OEM hardware manufacturers, but a bitter pill for companies like IBM.

Have no doubt – these companies wouldn’t be cozying up to an upstart like Pivotal, and having to swallow the bitter pill of playing second fiddle to the newbie, unless they had to. Desperate times call for desperate measures and, quite frankly, the rise of AWS is this generation’s burning platform – legacy vendors have a very limited time span within which they can react. An open source PaaS is a good option. So long as it remains a relatively level playing field – and this is were an independent foundation would come in.

I suspect that a more independent entity to run the Cloud Foundry project will become more and more important – already I’m hearing murmurings of concern within the community about Pivotal’s intentions, and how its relationship with community members will change as it attempts to balance being the instigator and largest single contributor to the project with its need to derive meaningful revenue from the project. It’s an age-old issue and one which really can only be resolves by a strong separation of the open source governance and the commercial aspects of the project. I put this to James Watters, heads of Pivotal’s Cloud Foundry team. His response was that:

We will be sharing some more thoughts on additional CAB members in the next few weeks. We are working on ways to ensure that its neither about the size of your company or tenure by default, but rather the demonstrated leadership within the community by contribution, use and collaboration.

That comment speaks to the tensions that must be felt within Pivotal – it’s always tempting to drop  the smaller partners and concentrate the balance of power with the higher profile ones – I’m sure Pivotal must be wondering the relative attention benefit of having an IBM, versus a Tier3 in the ecosystem. They’re also well aware of the tensions that have been rife in the OpenStack ecosystem – and will be working hard to ensure that Cloud Foundry manages to maintain credibility by not being dominated by two or three big vendors. Lots of unknowns, and it’s certainly going to be interesting to watch the progress on this one. In the meantime I imagine we’ll see more companies added to the advisory board and over time that will likely develop into something grander.

6 Comments
  • Well, I agree with Mr. Kepes that PaaS is the sleeping giant of cloud computing as everything will tend toward increased abstraction over time. I also agree that the incumbent hardware vendors whose badges have emblazoned generations of servers need to think hard and fast about their futures in a cloud computing world where commodity hardware is the norm. Hopefully, Cloud Foundry will learn from OpenStack about what not to do and hand over governance to a foundation not dominated by huge vendors with sharp elbows and outsized egos.

  • To say that OpenStack isn’t dominated by huge vendors with sharp elbows and outsized egos is patently absurd.

    As far as I’m concerned OpenStack will fail to live up to its potential precisely because of the dynamics of its governance.

    There are successful open source projects and businesses ecosystems that don’t center on a foundation and plenty of abandonware and dysfunction inside of the foundation umbrellas.

    stewardship > governance

    If Cloud Foundry is not being managed in a way that cultivates a community, the code is Apache 2 and any individual or coalition is free to make whatever choice they require to proceed.

    If anything, OpenStack has suffered from less than stellar stewardship and the dynamics of the foundation contribute to that so don’t hold that up as a shining example of how to do things.

    A foundation is neither necessary nor sufficient for stewardship.

    That is what these projects really need and they will only have utility to the extent that they do, foundation or not.

    0.02

    Sincerely,
    A nobody who isn’t in the Top 100 Cloud Computing Experts on Twitter.

    • Andrew – Openstack for sure has some big companies and sharp elbows but sunshine, as they say, is the best disinfectant. I love the idea of a CAB, but advice is very different from governance and I predict that the community (and the customers concerned about long term changes) will want the extra security that an independent foundation would bring..

      • What is the effect of this governance?

        What would actually be different?

        What are you claiming needs disinfecting?

        How does a pay to play seat on a foundation provide sunshine?

        What extra security does a foundation provide?

        Are you going to ask Redhat to put all their stuff in a foundation?

        What about Opscode or Puppet Labs?

        I can make good arguments why code should belong to a foundation, but nebulous ‘governance’ isn’t one of them.

    • Oh and don’t be angry re the twitter list, you and James are kind of the same thing and he’s there so be default you make the grade

  • Pingback: Is Cloud Foundry the Next OpenStack? | The Diversity Blog – SaaS, Cloud & Business Strategy | Dan Gorman's Technology News

  • Pingback: Who's Who in Cloud August 16, 2013 Gathering Clouds

  • Pingback: Is Cloud Foundry the Next OpenStack? - Jan Brass

Leave a Reply