The Utility Computing Myth : The Real Reason Why The Sight Of A Fedex Truck Makes Me Giddy

So here’s the crux of the utlity computing mantra: all that processing power and hard drive space on your desk isn’t needed for a great majority of applications. Instead, we’ll have big, huge datacenters serve up applications over the web that help you get your work done. You’ll pay either a subscription fee or stare at ads to subsidize their existence. No more $199 for a box with a CD in it.

The days of downloading, unzipping, installing and and then "loading" software are numbered. Instead, we simply visit a URL and bam…instant application and away we go. So watch out Microsoft and Intel, we don’t need kick ass PC’s anymore. Watch out Seagate and Maxtor, we don’t need your hard drives anymore. We just need a thin client good enough to run a browser. The revolution has begun. Right?


Anyone that’s messed with web-based word processors like Zoho Writer or Google Documents knows that they’re nowhere near as impressive as their desktop counterparts like Microsoft Word. They lack the responsiveness and fidelity and are short of a multitude of features. We’re ok with it in some cases because it’s damn convenient,

Still, web applications have come a really long way. They’re getting better and better. If you want to see the future, check out Adobe’s Buzzword. It’s a very impressive, Flash-based word processor that gives anything you’d install on your desktop a run for its money and it requires you to just visit a URL. So maybe this is the revolution, right?


Let me put this in a not-so-subtle way: regardless of how you got your software, software still needs to be impressive and competitive. That datacenter is doing two distinct things, and neither are very revolutionary (in my opinion):

1. From Pony Express To Fed-Ex Next Day

fedex_truck First, they’re eliminating all the Old World steps that were once necessary for delivering software to you. CD’s. Registration Keys. Big, huge installable downloads. All those features don’t need to be sent in one big crate anymore. The prerequisites to using software are virtually gone. Just go to a browser, wait a few seconds and off you go. This is how Youtube took over the world and Real Player died an ugly death.

2. Real Estate For Your Bytes

10LSA_pop Second, the datacenter will house the assets we create with these tools. Images. Documents. Spreadsheets. They’re all stored centrally and available to me wherever I am. This opens up new opportunities for collaboration and sharing and of course, it’s very convenient. There’s still an important privacy barrier but let’s assume privacy is a non-issue for the sake of this post.

I Don’t Need A Mult-Billion Dollar Datacenter In North Carolina To Resize An Image

For whatever reason, everyone is equating this with the end of traditional software as we know it. As far as I see it, that CPU on your desk is needed more than ever because it’s still doing all the work. Yes, the software is being delivered differently, but the processing is still happening locally. Yes, Buzzword fires up in your browser, but it’s your computer that’s doing all the magic. In fact, one could argue we need faster processors on our PC’s because so much of today’s software is interpreted on the fly.

The Adobe AIR platform is all about delivering desktop convenience through a seamless (or near seamless) install experience. AIR is about leveraging what your PC can do, not what some monolithic datacenter in Indiana can do. Of course, your data can still be stored centrally and there’s no longer a need to go through cumbersome installations, so the advantages are clearly there.

wall_outlet Many have argued that this notion of utility computing will transform the software industry. Computing power will be metered out like electricity as pay-as-you-go infrastructures take over. Undoubtedly, this shift will have an impact on business infrastructures. Small businesses will no longer need to host mail servers and web applications. They still have to make platform choices and still have to either buy, build or subscribe to software, but the IT headaches, to a large extent, will be outsourced.

Back To Square One

Even so, you still need and want great software. It’s important to distinguish the "utility provider" role that Google plays for example (providing storage space, support, uptime, high-bandwidth) with the "software provider" role that Google plays (providing spreadsheets, word processors and email). Amazon, Google and Microsoft are going to provide the utility infrastructure. That’s a given. As a result, the "utility" end of the value proposition will quickly commoditize, if it hasn’t already done so. The differentiator lies on the "software" end of things – in effect where its always been.

aufmacherRegardless of where your software originates and its means of delivery, if it’s better than a competitors, it will win. This key fact has not and will not change. And to win, you will still need to take advantage of the power on people’s desktops and laptops. Technologies like Silverlight and Adobe AIR reaffirm the need to marry a click-and-run experience with the richness and power of desktop software. But more importantly, they validate the importance of delivering better software than the next guy, however way it gets there.

We may well pay for hosted software like we do electricity today. But the real variety of experience and innovation will not lie in the power outlets, but rather what we plug into them.

3 Comments The Utility Computing Myth : The Real Reason Why The Sight Of A Fedex Truck Makes Me Giddy

  1. John@PM

    An underlying assumption from your post seems to be that the user will be using a single machine. From a corporate-software-decision-making standpoint it makes lots of sense, but when we consider apps for individuals like googles suite or the adobe products, I feel the real benefit is mobility. I agree the primary benefit of externally hosted apps isn’t a lightened load for the hardware, but also that the freedom for the user to access the app from anywhere with a browser is a significant draw.
    As a student over the past few years, I’ve often worked at four or more separate machines during the day, but still access almost all of my software and documents from the web or via novel servers. I am more and more willing to leave the house only carrying my smartphone rather than lugging my macbook around because I know many of the services I use are up in the proverbial cloud somewhere. I still need local horsepower, but I don’t need to own that hardware necessarily.

  2. James Urquhart

    John: I would go a step further and say that from a corporate-software-decision-making standpoint the variety of devices (aka edge devices) each indivdual user or group of users accesses is an issue in how they select software moving forward. This idea that a single PC will hold my world is…well, not exactly dying, but certainly morphing quickly.

    Ask yourself how much of your daily deliverables are primarily stored off of your PC. I will grant that it may vary from person to person or company to company, but in my personal life more and more information is being stored “in the cloud”. (For example, I’ve gone “paperless” with my bank; now all statements are their problem until I need one.) This also seems to be true for a majority of the technology professionals that I work with.

    My wife (a student) still owns and uses a laptop. However, when we bought it, it was cheap and minimally powered. As more and more of her world gets stored “in the cloud”, the need to increase horsepower diminishes, to the point that her two year old laptop will probably last her another 2-4 years (assuming the cheap Gateway parts last that long).

    Utility computing probably has more to do with server-side economics than client-side, but as bandwidth grows and cheapens, the quality and power of online software grows, and the marketplace for cloud services matures, the economics of locally installed software begins to diminish rapidly.


    I think it’s the same cyclic sort of rotation we saw 10 years ago, when Sun promised that desktop computers would “die”.
    One thing I disagree with is that rich client application UIs are superior to web-based UIs. Google Maps is a great example of this. Atwood’s got a good post:
    It may be a case where the very best web apps are already outperforming their desktop equivalents, whereas run of the mill web apps are still crap. Then again, so is most desktop software.


Leave a Reply

Your email address will not be published. Required fields are marked *