BlackBerry outage: How much reliability is necessary?

Computerworld - How much reliability can you afford?

Last Monday afternoon, BlackBerries stopped working again. For about three hours, the CrackBerry addicts couldn't get their mobile e-mail fix. The root cause: an unsuccessful infrastructure upgrade by BlackBerry vendor Research In Motion.

This outage came less than a year after an 11-hour BlackBerry service outage last April. Root cause that time: a different infrastructure upgrade.

Naturally, BlackBerry users were unhappy. A few industry blowhards said they were furious and grumbled darkly about a backlash against the BlackBerry.

It won't happen. And not just because of the BlackBerry's supposedly iconic status or because users have some irrational affection for RIM.

The analysis is simple: The BlackBerry works, mostly. It's cheap, functional and -- most of the time -- reliable.

Three hours down out of 7,000 since last April? That's 99.95% reliability. Want more nines? You'll have to pay more.

Probably a lot more.

That's because most users don't need all those nines. They can survive a few hours without e-mail arriving in their pockets. Before the CrackBerry, they did it all the time.

So only a small fraction of BlackBerry users would ante up for service with much higher reliability. That would likely make the cost per user stratospheric.

Cheap, functional, reliable. You get to pick two -- at most.

The definition of "mission-critical" is that we choose reliable over cheap. For anything that's truly mission-critical, we spend the money. That's why we have backup generators and redundant T1 lines -- because we know drunk drivers plow into electrical substations, trucks skid into telephone poles, and backhoes and boat anchors take out fiber-optic cables.

So by definition, BlackBerries aren't mission-critical. They probably never will be. There just aren't enough users willing to pay for the battle-hardened, fully redundant BlackBerry infrastructure that more nines would require.

Ah -- but how much reliability can you afford?

You can't guarantee that every e-mail message will go through. But maybe you can afford to pay for a return receipt for every message one of your users sends. At least then they'd know if the mail was going through or the system was down.

Or what about a fail-over system that lets your users collect their mail some other way on those rare days the BlackBerry dies? Web mail by phone, e-mail converted to text messages, iPhones for everyone -- there are lots of possibilities. But they'll all cost extra. How much can you spring for? How much should you spring for?

That's defined by how much reliability your users actually need. And you can't know that without asking them.

And right now, you have the perfect opportunity to do just that. They've just been through a BlackBerry outage. It's been a week, so it's still fresh in their minds, but distant enough that any minor sting has faded.

So ask them: Did it cripple their work? Was it an annoyance? Did they even notice? How much of a work-around would they be willing to use to get e-mail another way next time? How much would they be willing to pay?

Don't hint at solutions or imply that you've got a magic bullet. Just ask and listen. There's a good chance they'll tell you just how much functionality and cheapness they're willing to trade off to be sure of getting their e-mail on the go.

And that's how much reliability you can afford.

Frank Hayes is Computerworld's senior news columnist. Contact him at frank_hayes@ computerworld.com.

Read more about Management and Careers in Computerworld's Management and Careers Topic Center.