Business IT hates single points of failure. Hate's 'em. With good reason, too, since a single point of failure means that some significant portion of your operation is at the mercy of a single component or vendor. One of the most common and critical "single point of failure" locations is the provider for your wireless network. It's seen as an unavoidable side-effect of the way we do wireless business in North American, but it can still leave IT in an uncomfortable position when it comes to keeping wireless applications up and running. Last week, corporate customers (and individuals) were reminded of the risks that come with a single point of failure when AT&T suffered a massive network outage in San Francisco. The outage hit in the middle of a busy afternoon and was quickly discussed on-line as enraged users took to Twitter to share their anger.
So what do you do about the problem? One aspect of the solution is to be vigilent about the decision of which wireless provider to choose, though even that can be difficult. Take AT&T, for example: InfoWorld's Cringely says that AT&T is strictly for losers, while respected users like Instapundit's Glenn Reynolds brings more data to the party (though it may not make the picture much clearer). Over at the New York Times, Randall Stross finds data that indicates that AT&T may be just as good as Verizon -- it's all the iPhone's fault.
So what do you do with all this information? One of the primary tools used in wireline networking -- multiple providers on separate links -- isn't really an option for wireless networking. Or is it?
In most cases, multiple wireless network providers isn't going to be an option for any single smart phone. Even if we reach the point at which any phone will link to any network, the phones aren't going to connect to more than one network at a time. That means there are really two options: scatter or diversify.
The scatter option is the side-effect of companies allowing (or requiring) employees to bring their own cell phones. Since it's highly unlikely that all the employees will choose the same phone and carrier, the enterprise ends up with carrier scattering by default. This can mean great things for the organization, though it won't help any single employee very much. With enough scattering, the odds are good that the organization will be able to keep functioning even if one carrier suffers a wide-spread outage. The IT professionals I've talked with aren't using this as a justification for the employee-owned phone option, but most are happy enough to accept it as a pleasant side-effect.
Diversity comes from the same ability that allowed hacked-off AT&T customers to tweet their displeasure: WiFi capability in the cell phone. Right now, only a few phones have the ability to move their communications to WiFi, but it's going to become a matter of growing importance on check lists, with benefit to both the carriers and the customers. The carriers get to enjoy reduced network load from data-hungry customers, while the customers get some level of protection from network outages. It's not a perfect solution, but it's a real step in the right direction, especially since it's still an option for companies that want to standardize on a single phone or carrier for all their mobility needs.
Now, if you're going to move to the WiFi option, there are still issues you need to take into account. One of the most important is daily fees for WiFi access (Starbucks, anyone?) though there are decent options for this. I've written about Boingo in other venues, but I continue to be pleased with the flexibility they offer when I travel. A single monthly fee gives me access to WiFi hotspots at Starbucks, most airports, many restaurants, and lots of other places. It's a good service, and the kind of problem-solver that enterprise IT folks will need to lean on to make their wireless network more reliable.
Occasional outages are almost inevitable -- the important thing is to plan for them, rather than trust to luck to keep your apps safely working.