The Blog

  • April 1, 2010
  • Put a stake in the ground for Cloud Computing

    Got this in a newsletter from Ping Identity Corporation (pingidentity.com) the other day and it made me think.

    “A few weeks ago, we held one of our quarterly customer advisory meetings in our South East and Mid-Atlantic region in the U.S. These meetings always provide insight into the evolution at our large enterprise customers. Here are some of the themes that emerged:

    “Cloud” is defined as anything hosted (SaaS, PaaS, IaaS, BPO, ASP) and the flood gates are opening.

    Enterprises want a loosely coupled architecture, or in other words, no single vendor and integration is a key requirement.

    Really, I thought?  If Cloud is defined that way, why have another term?  Why don’t we just keep hosted and save everyone the trouble of spring cleaning our memories?  Out with one term and in with another!

    I am not sure if this statement was simple reporting from what Ping discovered from their users or if it’s an attempt to make a definition.  If it’s the former, I can understand well enough because at the point of paradigm change we always see a degeneration of terminology into mush.  The new term is adopted to cover the old paradigm long enough for everyone to jump into a lifeboat headed in the general direction of the new idea.  If that’s it, then that’s that.

    But I get queasy when I think that some people in positions of decision-making authority might actually believe it.  I don’t know about BPO as an idea but ASP is a retread from the Clinton administration and I thought it died when USi failed to secure customers.

    Cloud is important and it shouldn’t be trivialized or homogenized into something that’s just like everything else.  The reminds me of 8 year-old soccer leagues where everyone gets a trophy for showing up.  But we’re all big people now (mostly) and this ain’t that.

    Cloud is a new way of computing that accesses resources from the Internet or “Cloud” and creates orders of magnitude better economies for users.  It is a necessary component of ubiquitous computing.  Many, many companies can’t do cloud or Cloud yet and, using the lifeboat analogy, they’re jumping in with what they have.

    You can imitate Clouds and many vendors are doing this but imperfectly.  Think of the IT maxim — “Better, faster, cheaper, pick two” — and you’ll have an idea of what’s happening with vendors offering one or two components of Cloud Computing.

    To be clear, you need IaaS, PaaS and SaaS to do cloud computing and they all need to work together for the same reason you need hardware and software.  But even more importantly, this trio provides benefits equivalent to 1+1+1=4.  The four comes not from loosely coupled architecture but from tight coupling because tight coupling makes it possible for users to load and go without the traditional issues of integration.

    Loose coupling is not a customer benefit as I see it, it is a vendor benefit.  The coupling issue is causing heartburn for some people who make their livings making independent point solutions and performing integrations.  Tight coupling makes their business models less viable.  But they, too, will jump on a lifeboat and figure out how to play in a tightly coupled world and many already have.

    The one positive that comes out of all this is that if paradigm change drives terminology debasement, then I would say we’re about half way through the process.

    Published: 14 years ago


    Discussion

    • April 2nd, 2010 at 3:25 pm    

      Thanks for highlighting integration as a key requirement Denis! On that topic, here are my 2010 Cloud Integration predictions…which seem to be coming true. (Guess I should have gone out on a bit more of a limb!)

      http://blogs.informatica.com/perspectives/index.php/2009/12/21/2010-cloud-integration-predictions/

      Darren

    • April 1st, 2010 at 1:44 pm    

      You are right about vendor lock but that’s a phase I believe we all need to go through. Think of it like the operating systems wars of the mini-computer era. What ran on DEC did not run on DG because of OS differences and compilers. But customers got tired of that which gave an opening to Unix. I don’t think it will take long for customers to tire of vendor lock in and I see them trying to actually avoid it by supporting interoperability where possible. This leaves the ISV who is locked into a vendor when he builds an app on, say, the AppExchange in Force.com. It is to everyone’s advantage to keep the ISV happy because of the network effect — Force.com is only valuable if many people use it and the ISVs won’t use it if they are not treated well. Rather than vendor lock in, I like to think of this as the software industry’s version of a Mexican Standoff. To be sure it is a metastable state but it is one that could exist for a long time, balanced as it is.

    • April 1st, 2010 at 1:29 pm    

      Great post. Your definition of the “Cloud” term makes sense to me. Although in a scenario where the components for configuration/ development are tightly coupled with the platform and infrastructure…seems the pundits would say that creates vendor lock-in. I tend to think that’s an inherent trade-off given the overall benefits provided by true cloud. I’d be interested in your thoughts.

    Speak Up

    You must be logged in to post a comment.