Beware The Utility Pole

by Scott Kantner, November 2nd, 2011 in CTO

Winter served early notice this past weekend in the Northeast. The official family measuring device, an heirloom yardstick with the name of my father’s business on it, recorded close to 10 inches of October snow before utility power failed.


Within an hour, our broadband connection followed suit, after which we arrived back to my mother’s future – a world with no electricity or telecommunications. It’s actually a very quiet and peaceful place, but gets old quickly. Firing up our backup infrastructure catapulted us back to the 21st century with heat, water and light, but with no data service and little I could do about it. This, faithful reader, is the Achilles Heel of cloud computing.

In the next 3-5 years, cloud technology will mature to the point where its adoption will be mainstream, starting with private cloud, then hybrid, and finally the public model. But a vexing question will likely remain unanswered: What happens when connectivity to the cloud fails?

Enterprise business has several options for building multiple on-ramps to the Internet turnpike. They can contract directly with multiple carriers to install multiple circuits with diverse paths. They can install high-speed wireless on their rooftops. They can install broadband connections from one or more local cable or telephone companies. With the exception of the last option, most small businesses can’t keep up with their well-heeled enterprise brethren because of cost and sometimes, location. In the final analysis, however, it really doesn’t matter, as both have the same problem. Regardless of the size of your wallet, the best you can hope to do is control the ends of a cloud connection. You can beef up your end, you can pick good providers at the other end, but you cannot control the so-called, but rarely talked about “middle mile.”

And where exactly is the middle mile of a cloud connection? Sadly, you have no easy way of truly knowing, but one thing can be known with certainty: the middle mile is a complex spider web riding on an overhead aerial infrastructure that is either rotting from age right off of the poles, or being suffocated by overgrown trees. The connections we’re growing ever more dependent upon are increasingly susceptible to decay, falling branches and trees, and, as we saw earlier this past weekend, heavy pelting snow. Not to mention the occasional fire or train wreck in the wrong place.

Internet connectivity has not yet reached the dependability level we once enjoyed with our national power distribution system, nor is it likely to anytime in the near future. I say “once enjoyed” because clearly, power dependability is now in decline in the US, not because the weather is more severe, but because the system of aging wooden poles and transformers is not being effectively maintained. Take a good hard look at the poles, especially the cross arms, in your neighborhood and where you travel (not while driving, of course). The situation is, in a word, scarey. Or as my civil engineering friends would say, worrisome.

The upshot is that users of cloud services are in some degree putting their faith in 50+ year-old wooden poles as much as they are in their provider’s cloud infrastructure. In this light, the prospect of running all or parts of your business on SaaS offerings may no longer appear as bright and shiny as it once did. My purpose, however, is not to frighten you away from cloud. Rather, I simply want to awaken (or re-awaken) your awareness to the middle mile issue so you will think through your options and take the necessary steps to manage the risk.

If you’ve already committed to a full-blown SaaS strategy (e.g. SalesForce or NetSuite), your answer may be as simple as setting up some cell phones with tethering and sending your employees out with their laptops to find a working 3G or Wi-Fi connection when disaster strikes. This might be referred to as “The Starbucks Option.” Don’t forget to consider mobile printing capability, if necessary.

If you’re considering IaaS or PaaS, using multiple clouds or geo diverse data centers within a cloud will help lessen your risk to middle mile failures.

For businesses that already have an IT shop and are still in the cloud planning phase, the best solution may be to simply keep key processing capability in-house, and use the cloud for recovery and peak demand purposes.

Severe weather wreaks havoc on the aging and decaying middle mile with extreme prejudice. And, it’s not just one pole you need to worry about, but potentially dozens, so it all comes down to this question: Can your business afford to be down when the middle mile of utility poles is being resurrected from the dead, and how long can you wait? If not, what are your plans?

Comments