Blog post

Cloud Infrastructure: The Next Fat Dumb and Happy Pipe?

By Daryl Plummer | February 10, 2009 | 2 Comments

So Cloud is about infrastructure, you say?

I am always amazed at how fast new concepts or topics get pigeon-holed into the same old discussions we have had for decades. Now it’s Cloud Computing’s turn. Since dealing with new things can be disruptive, IT organizations, and people in general, always seem to be looking for something they can recognized before doing anything new. In the cloud world, this has spawned multitudes of discussions about Private clouds and infrastructure as a service and cloud platforms to replace on-premises platforms. Good stuff – as far as it goes.

What gets to me is when people stop there they don’t seem to realize that they are placing themselves on the same treadmill that the Telco providers have been on for those same decades. Problem is, the treadmill is moving faster backward than you can walk forward. Cloud infrastructure is a means to an end, not an end in itself. Sure you can get great value by either adopting or selling cloud infrastructure for next-generation applications. But, there are diminishing returns on that investment. If you are a provider of this infrastructure either in the public sense or in the private sense, you may find that your ability to use it to differentiate erodes faster than sand on a Florida beach. As a vendor of cloud infrastructure you must be afraid of what happens when IBM really wakes up alongside HP and a host of others – some of them new. Opsource competes with Amazon, competes with IBM, competes with Microsoft – all for the same dollar to host cloud infrastructure for businesses that have already determined that owning infrastructure is not usually a differentiating value proposition for them.

So, how many phone companies do we need – figuratively speaking? How many cloud Infrastructure providers will there be room to support? I say a lot at first – then a rapid decline in a bout 5 to 7 years. If you are still in this game at the level of infrastructure by then, you better have Google, IBM, or Amazon stapled to your butt.

So what is the answer for making sure you do not get commoditized into just a pipe (or platform)? It’s simple – get closer to the customer in the value chain!! When you are an infrastructure business in a commoditizing paradigm like cloud computing, you don’t want to be at one end of the pipe while your customers are at the other. And, you don’t want to be one of hundreds of vendors who say their customers are the ones who serve customers. That is a recipe for a repeat of the Telco dilemma. Oh, and the telcos are trying to get in the game too.

But the ones who have the real advantage are those who will either hold the hand of the customer (consulting, or who add value on top of the infrastructure services (e.g. more performance or security or additional services). In addition, there is a lot of room for sourcing companies to turn to cloud as a way of continuing to be important. Business process outsourcing is an opportunity. And what of systems integrators? Well, there is an opportunity here too. When people need to move data between services or  to migrate entirely from one service to another, the new breed of systems integrators should be ready.

Ok. These are just musings on the dilemma of changing market needs and approaches but it is something to think about.

The Gartner Blog Network provides an opportunity for Gartner analysts to test ideas and move research forward. Because the content posted by Gartner analysts on this site does not undergo our standard editorial review, all comments or opinions expressed hereunder are those of the individual contributors and do not represent the views of Gartner, Inc. or its management.

Comments are closed


  • Russ Castronovo says:


    Can you clarify a bit more how there are diminishing returns on the great value one can get by either adopting or selling cloud infrastructure for next-generation applications?

    I’m not certain I follow what you mean there.


  • Daryl Plummer says:

    Diminishing retuerns are elated to the idea that if you do private cloud computing then your opportunities to gain cloud value are limited. For example, elasticity can only go as far as the capacity you have already bought and have in your private cloud. In the public cloud context, elasticity is ever changing and opportunities are widely varied. Not so in the Enterprise. so, even though you get more flexibility, you begin to lose out as you need to continue increasing capacity with depreciable assets on site. Anoher example is economies of scale. Sure, you get them with a private cloud but they are miniscule by comparison to what Amazon or OpSource will give you. You can’t keep up with their ability to take advantages of economies. As time goes on, your customer base stabilizes and your economies of scale just become a capacity management issue again. Public cloud proviers will have a business model based on gaining millions of customers. I doubt most enterprises can hope for that level of scale. Also, remember that as the infrastructure commoditizes you are still held back on platforms that probalby have not kept up with technology and certainly are not as widely used as the public cloud platforms. Your returns on investment begin to look more and more like just a large heavily virtualized Data Center. One more example, although there are many more. In a private cloud, the services you deliver will be limited in number, in variety, in price, and in breadth of capability. Certainly, you can build and deliver the services you feel you need but how agile will you be in allowing your business to source new services or to swap out older ones? The more of them you are deliovering in a private context, the less capability you will have to mae changes to them or to add new ones since your resources are limited. A public cloud provider has resources limited only by the willingness of customers to buy their stuff. Additionally, there will be multiple cloud providersfrom whom you could choose. If you do private cloud, you are the only one. So, sure, use it, but remember you will also need public cloud or your value proposition is limited. And if you use pubic cloud in addition to private cloud, why not go for the whole thing except for really critical workloads – which probably should not be cloud yet anyway – private or otherwise. Make sense?