Flexibility vs. scalability in the cloud — there’s no easy answer

Dear Bob …

In case you missed it, the April 2011 issue of CIO magazine is all about cloud computing. I’m basically in agreement with your take on the cloud, but I’m worried that people who might be, umm, less familiar with it (CIOs) may now have even more fodder for pushing it into places where it doesn’t belong.

[ Want to cash in on your IT experiences? InfoWorld is looking for stories of an amazing or amusing IT adventure, lesson learned, or tales from the trenches. Send your story to offtherecord@infoworld.com. If we publish it, we’ll keep you anonymous and send you a $50 American Express gift cheque. ]

If you ever need a cheap joke, I’d check out the photo on page 16 — it includes the caption: “Lojas Renner CIO Leandro Balbinot says the key to successful cloud implementation is giving oneself over to the vendor’s processes.” (I’m not making that up.)

Please please please continue to push some logic against the “all cloud, all the time” mantra that it coming from other media outlets.

One of the main arguments for cloud computing seems to center on capital expenditures versus operating expenses (capex vs. opex), that operating expenses are better since it lets you be more flexible. I get it — mostly.

But I’d like to hear your thoughts. It seems to me that if I have 100TB of storage today, then I’m about 99.99 percent certain I’ll need at least 100TB next year and my capex investment is likely to be fully realized. Isn’t that a pretty good argument for capex? If not, doesn’t that suggest we should all be leasing cars instead of buying them? OK, maintenance costs make that a bad example, but you get the point.

– Cloudy

Dear Cloudy …

I’m flattered you think I have this kind of impact on the industry, but I’m sad to say my influence is far more limited. For whatever it’s worth: The quote is unfortunate. Somewhere inside it is a valid notion — that when you outsource any service, including hosting, an important part of what you’re buying are the vendor’s processes and practices. I’d say that’s a matter requiring serious due diligence, though. This is a business relationship, and “giving oneself over” has connotations I find unnerving on multiple levels.

On the opex/capex issue, I’m pleased that industry commentators are getting this right — well, half right. It’s like this: Opex means lower fixed costs and higher marginal costs, while capex means higher fixed costs and lower marginal costs.

Label the high-fixed/low-variable cost approach “scalable” (in the economic sense) because as volume increases, costs don’t climb very much. The downside is loss of flexibility (also in the economic sense); if volume decreases, your costs don’t go down very much. Most of them are, after all, fixed.

The opex approach earns the label “flexible.” As volume shifts up and down, you can add and shed costs more or less on demand.

What the commentators miss — possibly because it isn’t in the press releases so many seem to rely on — is that neither flexibility nor scalability are good or bad, other than in the context of your volume forecasts. If you expect steady growth, scalable — therefore, capex and the use of internal infrastructure — fits your situation. If you expect volume to be highly volatile, you should go for flexible, which means the cloud might be the ticket.

Two factors complicate this analysis even further. The first is the economy of scale, which affect small companies considering the cloud as an alternative to internal infrastructure. Because small companies don’t enjoy economies of scale, cloud vendors can spend less capex per unit of computing resources and pass along some of the savings. The result: For small companies, cloud-based solutions can also be more scalable than internal infrastructure.

Meanwhile, especially for larger companies, capex can improve flexibility, namely through the use of “private clouds” (which really should be called “fogs” — after all, that’s where you find yourself when you’re in the middle of a cloud). That’s because of how IT has had to manage capacity in the past: by provisioning enough spare computing capacity to support each application separately.

These so-called fogs allow companies to consolidate spare computing capacity to single pools of storage and CPU cycles, reducing the total amount of spare capacity needed. It’s using capex to increase flexibility.

Something else strikes me about the capex/opex argument: It’s a terrific example of the “first liar wins” rule. It doesn’t make actual sense, but no matter — it’s argument by assertion, which can be very effective so long as you’re the first one to get your statement in front of everyone.

– Bob

This story, “Flexibility vs. scalability in the cloud — there’s no easy answer,” was originally published at InfoWorld.com. Read more of Bob Lewis’s Advice Line blog on InfoWorld.com. For the latest business technology news, follow InfoWorld.com on Twitter.