14 Comments

Summary:

The next big leap in both technology and business models around sharing elastic compute resources will be bidding for those resources at auction or acquiring them through a broker, according to Forrester. But this broker business just adds more abstraction to an already abstract business.

istock_000004385975small

The next big leap in both technology and business models around sharing elastic compute resources might be bidding for those resources at auction or acquiring them through a broker, according to a report issued Thursday from Forrester. The report argues that while the technologies aren’t here today, a business built on aggregating clouds from public and private networks and delivering capacity and services in an automated, on-demand fashion to companies will emerge by the end of this year, and standards will start appearing in 2012.

Forrester believes that today we have a simple cloud broker model where companies can buy access to infrastructure as a service on demand, things such as Enomaly’s SpotCloud fit this mold. However, in the future, it expects a”full broker” to emerge that automatically spans all clouds and acts as a “fixer” for companies seeking to deploy new apps or handle busy times of year. The fixers don’t just do this for compute resources, but also will one day do it for people, making the idea of hiring seasonal workers something the full cloud broker could do — providing both the IT resources to support those people but also the people themselves.

Forrester's complex cloud broker model isn't for the faint of heart.

That’s a bit of a stretch for the cloud, seeing it as this universal resource provider, but the report takes us back to the earth’s stratosphere by pointing out that for this full broker model to develop, the tech industry will have to build some tools: namely, those that will make so-called “cloudbursting” a reality. Cloudbursting as a concept refers to moving data and applications instantly from one cloud to another. The number of reasons this doesn’t happen today begins with a lack of interoperability between clouds and ends with issues tied to moving virtual machines from one data center to another.

But once a vendor pulls all this together, it will be on the cutting edge of the next generation of cloud services and reap a “healthy 5-percent margin” in the process. If that doesn’t strike anyone as particularly healthy, consider that Forrester believes cloud computing margins will drop to commodity levels … eventually. From the report:

Although cloud computing is probably still two decades away from being a commoditized utility, the underlying delivery models will follow those of other utility business models. For example, high standardization in electricity grids has moved the entire industry away from flat, volume-based pricing for industrial consumers to the spot pricing of today’s smart grids.

My Take: I respect that Forrester is trying to predict the future here, and aspects of this model are clearly coming. However, it also posits that sophisticated private cloud builders will link their private clouds with public ones and even make their private clouds available. This could allow a broker to “offer a ‘virtual virtual’ machine at an even lower price point than a commoditized IaaS offering.” Forrester notes this would take deep trust, and I think it may not happen at all, especially since many sophisticated CIOs are also in heavily regulated industries.

The idea that sharing human resources as well as IT and linking them together is pretty bold prediction and would create a dichotomy in employment experience between professionals and middle-class call center or service workers that would have implications for society and public policy that boggle my mind. Anyhow, that’s the future as Forrester sees it — the cloud becomes more pervasive and more nebulous.

  1. Stacey, you are so polite and diplomatic. This prediction is nuts. Not gonna happen. Wouldn’t be prudent, not at any juncture.

    Yes, electricity can be commoditized, but electricity is the same everywhere. It’s a flow of electrons. Computing is so different, so obviously different that if it has to be explained to someone, it’s likely they won’t comprehend it.

    This is just more of the same BS being propagated by the cloud computing merchants, who want everyone to believe that choosing to outsource your computing is the conventional wisdom. It’s not, and will most likely never, be a good idea. The economies of scale offered by cloud computing vendors do not offset their overhead and profit margins, not to mention the cost of bandwidth between the users and the utilities, and the associated risks they bring to the party (which they will not take responsibility for, it’s in the fine print). The utility model works for electricity because until recently you couldn’t buy a safe, reliable power plant for your home the size of a refrigerator. But you can buy a cheap, small computer to meet all your needs.

    Share
  2. In the spring at the Midsize Enterprise Summit, Gartner predicted the rise of the cloud broker. I highlighted it in a blog posting at the time; http://turningtechinvisible.blogspot.com/2011/07/repost-gartner-vs-googles-cloud-one-of.html

    Share
  3. Cloud brokerage is a real trend– our enterprise customers look for a new type of partner to help them weave together cloud services into business solutions. My take on that here, in ComputerWorld: http://goo.gl/KXglF

    But what’s nuts about this prediction from Forrester is the expectations that anyone is going to be able to transparently shift workloads back and forth between private, hybrid, and public clouds… that requires a “lowest common denominator” approach to cloud computing that ignores all the innovation occurring at the PaaS and IaaS layer.

    So agree with KenG on his point re: commoditization, but disagree on his point re: outsourcing. ;)

    Share
    1. Completely agree. Fundamentally cloud is a service-oriented phenomenon where you are able to access specialised services externally to replace those you might traditionally operate in-house. The services that businesses are interested in are actually, uh, applications, process templates or complete outsourced business services, however and not infrastructure (which is essentially worthless). As a result I totally agree that the concept of a ‘broker’ is not that I can take complex, heterogeneous assets with many environmental dependencies and somehow ‘magically’ have them appear on a different infrastructure without losing all of their surrounding context but rather that I can access distributed business services in an integrated way to build extended value chains. To do this I need companies who will help me select, integrate and operate the complex web of services that I will require to operate a business in the future and to provide me with a single point of support, billing, experience etc. This much more valuable and business-oriented form of brokerage respects the diversity of services and their implementations and looks at integration at a much higher level of abstraction – avoiding the need to race towards the lowest common denominator of the “VM” for no purpose beyond a lack of imagination from the majority of technology companies and users.

      In fact one piece of advice on my blog recently was to:

      “Think of IaaS as a tax on delivering value outcomes and try not to let it distract you as people look to make it more complex for no reason (public/private/hybrid/cross hypervisor/VM management/cloud bursting/etc). It generates so much mental effort for so little business value”

      Share
  4. The prediction is spot-on and the name of the company is Charity Engine. We’ll have a trading floor for CPU time by the end of 2012, much like GridEcon or AWS spot pricing.

    Computing is like energy; the more you have, the more you can accomplish. And no matter how fast your own hardware, the sum total of everyone else’s hardware will always be far greater, scalable on demand and a lot cheaper to use.

    Faster, cheaper, on-demand. Anything that can run on a cloud, should run on the cloud. That’s probably why AWS is approaching a $1Bn turnover and everyone is happy to use the likes of Dropbox, Hotmail, Flickr, Google Apps, VMware, etc…

    Not gonna happen? Not gonna stop, more like.

    Share
  5. I find it interesting that the name for it is “cloud computing”. We associate clouds with cloudy days and not many people like cloudy days. This is just a connection I find interesting.

    Share
  6. Isn`t this interoperability fix what VMware have been trying to do with vCloud and vCloud Director? i.e. it is a reality already, albeit in early days.

    Share
  7. pretty good! my company have planned to establish a cloud computing center in a second-tier city of china PR. too much IT service providers thought there is no boundaries to cloud computing, in a sense, it do have no technical boundaries, but it do has saleservice boundaries. it is hard for a service provider to cover much more Accounter although its IDC has enough resource and capabilities. as an International corporation , how to make your China strategy is more important than technical application. in front of you , is a huge blue ocean market.

    Share
  8. Cloud computing sounds unsecure to enterprise ear’s… I am sorry to tell that, but this article and the nice graphs it is just another try to keep the wheel turning. A marketer with no idea just got a couple of concepts from a technician pushed to something “new” and ended maken a photoshop. Great job!!

    Share
  9. Cloud brokerage assumes that all the services are cloud agnostic. While in theory this could happen (if all adopt the same APIs and standards), this is still somewhat distant…
    At Xeround we offer a cloud agnostic Database-as-a-Service for MySQL based applications on Amazon, Rackspace and other will follow shortly.
    Unfortunately, even the pay-per-use pricing model is far from being supported; world is still behind with image-based service, tiered based pricing, let anone brokerage of anything, anywhere, anytime…

    Share
  10. Are sure this could really happen. what a twist in cloud computing trends this would be.

    Share
  11. Stacey,

    I would love to chat with your about cloud bursting and the fact that it is available today.

    enStratus has a cloud bursting solution including bursting between any of the 18 private and public clouds we support.

    For example, you can create an automated deployment and have first 100 VMs for your application begin in the private cloud and the 101st VM and beyond burst into the public cloud.

    Share
  12. There are so many caveats right now of even running simple applications (Like effectively running MS Sequel quick enough through a Cloud Server) that this is a long way off….

    Also with hacking on the rise, enterprise CIO’s are very weary of even running a Cloud Server. I hate to sound like a broken record… but this is the reality. It’s going to take some time for Private and Public Cloud Vendors to allow traffic to flow from each vendor… Right now it’s very restrictive…. Maybe if standards were made, this could happen but it is going to take a long time…

    Share
  13. Ofir Nachmani Sunday, August 14, 2011

    Great comments .. Those actually prove that ‘Cloud computing’ is now meaningless – http://www.infoworld.com/d/cloud-computing/its-official-cloud-computing-now-meaningless-169416 … Most of us talk about it as the IaaS layer some talk about it as the SaaS layer, the term is simply too general. I tend to say that the F report is pretty visionary and there is no doubt that there are no real standards in the market today.. But guys, hold on .. The cloud hype is almost behind us and we are heading to a world with standards .. You will probably ask what ? Why ? ..
    I have a simple explanation starting with Amazon AWS who started the IaaS market and is a head doing a great job with guiding all the others (just like Apple did with the iPhone) The brokers and web developers should make sure to be create their cloud layer so eventually they will be able to plug & play other IaaS vendor on demand.

    Following other specific “marketing comments” I will make my own.. Check “I Am OnDemand” blog – http://www.iamondemand.com

    Ofir.

    Share
  14. @kdierc It exists and is called cloudbursting. http://t.co/mSLRqDQu

    Share

Comments have been disabled for this post