15 Comments

Summary:

AppFog CEO Lucas Carlson isn’t shy about touting PaaS as the ideal way for developers to access cloud computing resources, but he also knows it’s not mainstream. In this inforgraphic illustrating the evolution of cloud computing, Carlson says PaaS will hit its stride in 2013.

appfog_infographic_013012

Lucas Carlson, CEO AppFog

AppFog Founder and CEO Lucas Carlson isn’t shy about touting platform-as-a-service as the ideal way for developers to access cloud computing resources, but he isn’t blind either. Although PaaS has been around for a couple years now and has already spurred hundreds of millions in M&A spending, Carlson knows it’s nowhere near the mainstream yet.

Carlson lays out his version of the evolution of cloud computing in the infographic below. Right now, API-based infrastructure-as-a-service offerings like that from Amazon Web Services and SysOps (or DevOps) tools are developers’ best friends in the cloud. Application-lifecycle platforms such as Cloud Foundry (the VMware-ran open source project  on which AppFog is built) and Red Hat’s OpenShift  are poised to reach critical mass in 2012, whereas so-called “NoOps” platforms such as AppFog and Heroku will reach that point in 2013.

During a recent phone call, Carlson told me PaaS is the model of the future, not the present, because only about 2 to 4 percent of developers — the ones on the cutting edge — are actually using it right now. “As interesting as PaaS is, the majority of developers … have some very real concerns that are holding them back from actually going forward,” Carlson said.

Aside from illustrating the evolution of cloud-development tools, Carlson said the infographic also aims to clearly delineate the different layers of the cloud stack, something he opined on in a December blog post. PaaS isn’t a feature of IaaS, he explained, but “a full reinvention from the ground up.” Every layer has to fully understand the layers below because they must manage them, but the user experience and the resulting increase in developer productivity are what make the service.

  1. Interesting post and like the “NoOps” slant. However what’s missing is evolving to next gen intelligent applications where PaaS 2.0 has “Data as a Service” component tightly integrated that can leverage IaaS if you architect it correctly. With that said check out http://www.fluidops.com/

    Share
  2. Sure, give a concept a new name then you can claim it’s only been around a few years. 10 years ago the concept was xSP and before that it probably had another name.

    Share
  3. GigaOM is running out of money so it has resorted to “infomercials” like this. Why don’t you openly put a label on this article ..”SPONSORED ADVERTISING”

    Share
  4. I love CloudFoundry and the PaaS concept to the point I’ve been building a product around it, but can we pretty please stop with this NoOps nonsense? The infographic is all about startups when the market clearly isn’t. There are so many realities, even startups that are no longer such, which can’t survive on a PaaS because they need specialised environments and this isn’t the exception, you can see it over and over in the market, especially for companies dealing with big data. Of course there are big advantages and it’s a blessing for people that want to experiment, but between that and picturing an IT world that no longer needs ops there is a big difference.

    Also I find a lot of the comments on cutting costs and productivity to be quite misleading. Costs aren’t cut, in fact if you look at the bandwidth cost or CPU/RAM cost per unit it is higher on a PaaS than it is on baremetal and for obvious reasons. What has been cut is the upfront cost, but that’s a different thing. Likewise it’s not that sysops or the cloud have made us more productive, that’s misleading, what they have done is reducing time to market.

    Bottom line, PaaSes are a blessing and I welcome them, they will no doubt help enabling a spur of new products that wouldn’t have otherwise probably seen the light of the day and certainly they will remove for the most part the necessity for IT staff. But that’s only true for certain products and certain companies at certain stages. By no mean it represents the IT Market and sysops/devops are here to stay and just get stronger as products get more complex and so do the infrastructures needed to support them.

    thanks,

    @spikelab

    Share
    1. From what I understand about what he’s trying to show with this infographic, the point isn’t that ops are going away, but they’re going away for developers. Someone still has to manage infrastructure, but that could happen anywhere down the stack. If you’re using a so-called “NoOps” platform internally — which will become fairly common, I believe — sysadmins will still be in-house, but developers will be free of that responsibility. It’s the evolution from physical to IaaS, which changed the roles of devs and admins, and then from IaaS to PaaS, which changes them again.

      Obviously, it’s open for debate, but I think that’s the point of the infographic. Check out the comment thread on Hacker News for some more nuance: http://news.ycombinator.com/item?id=3535657.

      Share
      1. hey Derrick,
        thanks for your reply.

        I don’t have any problem with the idea of self-serving, which is what I think you’re talking about, but that’s entirely different than what the infographic argues. I welcome PaaS as an evolution of self-serving and certainly you want to avoid having your developers waiting on your ops folks to be able to deliver value, but that’s far from saying that 2013 will be the year of NoOps. If, as you say, someone still has to manage the infrastructure ‘NoOps’ does not just make sense as a term, self-serving is a much better definition.

        But the even bigger problem here is the oversight of what PaaS represents in terms of market share. There’s no question it will grow and I very much hope so, but the same way IaaS has grown and hasn’t taken over, the PaaS won’t represent the end of sysops.

        The end-game isn’t NoOps, the end-game is faster time to market and the PaaS can help *some* of the companies out there with it. I think this distinction is very important and it’s not just getting hung up on words. Talking about NoOps is just creating attrition and misleading people toward a false goal. How would you feel if I told you the future is about making you irrelevant Vs helping me out moving faster and not depend on you so you can add values in other ways?

        I’ve expanded my previous answer on my blog here http://bit.ly/zWCqSh and I’ll follow up with another piece about NoOps and the communication problem behind it shortly. At the end of the day I think we agree, but it’s important to qualify some of those statements and work toward realistic and positive goals.

        cheers,

        @spikelab

        Share
  5. Derrick, Lucas,

    Great article and infographic. At Iron we have the same view of the future, but as some of the comments here and at Hn show, it’ll be a long road for mainstream to adopt.. particularly when there is pre-existing operational infrastructure (including people).

    And also the point is that Ops isn’t going away rather being “refactored” into specialized units: in this case the PaaS players.

    Chad Arimura, http://www.iron.io

    Share
  6. I have clarified some of the confusion about NoOps in our infographic in a blog post: http://blog.appfog.com/what-is-noops-anyhow/

    Share
    1. Hey Lucas,
      thanks for your clarification, it certainly helped. I’ve written some more thoughts that hopefully will helps us move toward the right direction. What you guys are doing is great, and we should figure a way out of those fights over terminology. More here: http://bit.ly/yAPE4o

      Share
  7. Definitely believe “PaaS” is the future way most cloud resources will be consumed. But today, PaaS pricing is too difficult to model out for systems operating at scale. Thus systems architects revert to running their own software on IaaS and drive down costs with DevOps automation. For more on this thinking checkout my blog post “The Problem with PaaS – Cost Uncertainty at Scale” – http://goo.gl/CHm7S

    Share
  8. I have a few comments on this. First, NoOps has actually been around for a while in numerous guises. Shared hosting is NoOps – you can build an app and deploy it without doing sysadmin. Second, the “exponential efficiency gains” are only realized if developers are sharing resources and are therefore limited on some elements of the system’s configuration. Otherwise, each environment has to be managed separately and it becomes and outsourcing/managed services model rather than a leveraged automation model.

    I think PaaS and NoOps is a great model for startups, as suggested by the infographic, because the developers can focus on building quickly and avoid operational hassles as they try to find their market. But in a scaling mode, the startup’s need to both develop profitable economics and add capabilities that may not fit the PaaS provider’s model means that many will have to leave this constrained environment for one with more control – thus bringing back the system administrator.

    451 Group talks about Auto-Ops – the idea of automating system admin. This is a related idea and it has its own set of strengths (more flexibility) and shortcomings (less granular scaling, still requires a systems architect somewhere in the process).

    So, while I think this is all good stuff, it really remains to be seen how important a role it will play. In some ways, it really is just a more scalable and reliable version of shared hosting – which is a huge market, but not appropriate for everything.

    Finally, I have a different reason than others for not liking the term “NoOps.” No-Op is a machine instruction in most CPU architectures that burns an instruction cycle with no effect (NOP in 8086 architecture). Thus, in some circles, the term “no-op” refers to someone or something useless. I don’t think this idea of leveraging a pre-configured ops platform is useless, so I don’t like the name.

    Share
  9. What interesting is that NoOps is growing everywhere, not just in PaaS land. For example, CircleCI.com (my company) saves developers from having to set up their own testing infrastructure. You can see the same trend in dozens of companies: hipchat, mailgun, airbrake, even Google Apps are all variations on the same NoOps theme.

    Share
  10. I would say that in a true PaaS world, NoOps is a little too ambitious of a state for the next 3-4 years. AppOps is a more realistic state to aspire for !
    Paddy Srinivasan
    http://www.opstera.com

    Share

Comments have been disabled for this post