On Cellular's (Non-)Reliability
What's the difference between 98% and 99.999%? More than three hours a week.
"That may not seem like a big difference but when calculated, it means 6.05 seconds in downtime per week for wired phones compared to 3.36 hours of downtime per week for wireless phones," write Adria Robinson, William Hunt, and Ryan Clough of the Samuelson Glushko Technology Law & Policy Clinic at the University of Colorado Law School. They rightly point out that the PSTN was originally built before the nation's power grid itself was completed, whereas the cellular network assumed a working power grid that, in the wake of natural disasters, may not be fully functional.
Justin Castillo has written about this issue before on No Jitter, and the need for regulators with different bailiwicks to work together to ensure that the systems they regulate, which are themselves interdependent, continue to operate in crisis situations to the greatest degree possible.
In the meantime, it's hard to know which systems will be working for which of your end users in any given episode of down time--whether a major natural disaster like Hurricane Sandy; a localized but still severe outage like those caused by Midwest flooding last week; or the many questions that swirl around the availability of cellular service during emergencies like the Boston Marathon bombing last week.
In that last example, news media widely reported that cellular networks in Boston were down, with some outlets claiming that the mobile networks were taken off line deliberately by the operators, out of fear that more bombs might still be hidden, and could be set off remotely via cell phone signal. The mobile operators denied that they had taken their networks off line, but it would be unsurprising if the cellular systems hadn't been overwhelmed by sheer volume of calls to and from concerned individuals.
In the case of a fairly straightforward local power outage due to severe but not catastrophic weather, end users with power over Ethernet might be able to stay on line in local offices, if the backup power there supports it. That was my experience during last week's flooding here in Chicago.
I'm not sure if any enterprise managers can really have a detailed knowledge of exactly how each local site will stay on line in case of a disaster or emergency of some sort. There are too many variables based on what services and applications the users use, what kind of event has occurred, and what sorts of backup resources are in place.
The Colorado Law School filers make a case that the PSTN remains the bulwark against outages that originate with the power grid, at least in events short of an overwhelming disaster like Sandy. "Consumers might not replace their wired landlines with wireless if they knew that wireless phones’ resiliency is 98%," they write.