The IT Road Less Traveled

by Thought Leaders at SwishData


The IT Road Less Traveled

Blog archive

The fishing problem for public cloud companies

By Jean-Paul Bergeaux
Chief Technology Officer, SwishData

Ever heard the story of the fisherman who was so focused on catching a big fish that he missed all the medium and small fish swimming around him? Many public cloud companies have hired high-paid sales people to go find enterprise customers. Get the big fish. It makes a splash in the news and makes the sales guy lots of money. But there’s a problem with this strategy. What about the easy-to-catch fish?

What is a public cloud really? It’s a modern virtualized data center paid for “as a service.” It’s really nothing technologically different. Amazon, Google and others offer the same thing that an internal modern virtualized data center could offer, but just as OPEX (operating costs) instead of CAPEX (capital costs). The theory is that the economies of scale of those public cloud companies will drive down costs that can be passed on to customers. I take issue with this theory for large enterprise customers.

There may or may not be some savings to be had for large enterprise customers in the public cloud, but if there is, it won’t be much. Here’s my version of the cost curve at a per-user basis:
 

Cost versus users in public clouds

That big dot in my opinion is about 2,500 users. So the cost per user is high when the user count is low, but is driven down as the user count grows and the economies of scale kick in. Granted, this is my theory. But in my experience in architecting solutions, at about 2,500 users, you have the ability to really use virtualization in a modern data center and anything above that is more about designing management controls, which don’t bear nearly the cost savings that virtualization achieves.

In fact, most data centers could see an uptick in cost as their user counts get higher and higher. Anyone will tell you that too big can become just as bad as too small. Why do we think for a minute that public cloud companies are going to avoid this while trying to manage mega-data centers? 

It’s crazy to focus on the big fish anyway. There are many smaller companies and organizations out there that will see tremendous cost savings and become hardcore believers in cloud computing as the adoption rate climbs. I looked up via the census the percentage of companies that have fewer than 2,500 employees and the number was a staggering 97 percent of all companies. That’s a cash cow for public cloud companies, and we haven’t even talked about individuals who don’t have to be employed to want to use public cloud services.

In fact, I think that it’s not only a mistake to not focus on smaller companies and government agencies, I think it is a mistake to even allow the larger IT organizations to become customers. It’s more likely that they will become disgruntled customers over time as they realize they aren’t saving as much as advertised and have lost control of their own IT, adding security and information assurance problems to boot. Why risk the public black eye like Google got when NOAA Cloud Computing Program Manager Stephan Leeb slammed them for not working with the agency to resolve ongoing security and information assurance issues on NOAA’s Gmail contract?

Want to hear more from SwishData ? Visit my Data Performance Blog, and follow me on Facebook and Twitter.

Posted by Jean-Paul Bergeaux on Jun 04, 2012 at 12:18 PM


Reader comments

Wed, Jun 20, 2012 Jean-Paul Bergeaux

I admit that this is just my thesis developed from years of solution architecture for large enterprises. I don't have numbers yet to prove it one way or another and the tipping point may be lower or higher than 2,500. I don't believe the tipping point approaches the 10,000 user mark. I'm working on possibly building some reference architectures, with cost, to begin to have a way to compare options with popular cloud options. No promises, but I am hoping we can create and publish one. If we are able to, it will take some time to build and vet before publishing.

Thu, Jun 7, 2012

2,500 as the tipping point. Interesting thesis. Would be very interesting to see some data behind that so that we can establish some empirical rules of the road. Thank you for your perspective. Enquiring minds need to know.

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Your Name:(optional)
Your Email:(optional)
Your Location:(optional)
Comment:
Please type the letters/numbers you see above