Withdrawal symptoms

BlackberryRIM’s outage this week, on the opening day of the Mobile World Congress, couldn’t have been timed any worse. Mind you, Alec Saunders reports that connectivity in Barcelona “has been an absolute bear.” A day later, RIM says that it is still investigating but the outage was caused by an upgrade to a routing system that had been recently upgraded.

I have been going back through my files and there seems to be a pattern of Blackberry network failures every 10 months or so. February 11, 2008, April 17, 2007, July 12, 2006, March 13, 2006.

We need to look at some trend analysis here. The July 12, 2006 outage was more localized and was blamed on a fibre cut. The other failures were more major network outages. What is going on?

Al Gore might blame the trouble on climate change, given the winter / spring time of year (February, March and April).

Is there anything to be found in the dates themselves? You will notice that the dates are in the teens. The 11th, 13th, 17th. Relevant? Maybe.

I might be more inclined to note that the failures were on Mondays and Tuesdays – given that software upgrades are usually a weekend activity, it may be that we are seeing problems that don’t manifest themselves until the stress of weekday heavy traffic loads are applied.

Why would the April 17th outage have waited until a Tuesday? That was ‘Tax Day’ in the US, and perhaps the traffic loads were even higher in the rush to file. Sure enough, the outage last April was due to inadequate testing of a new storage system.

I think most of us would generally like to see more communications from our communications service supplier, especially when there are problems. As end users, we don’t seem to get lots of information out of the folks at RIM, but I’d want to hear how they are going to improve their test plans and procedures before boinking with live service.

These are somewhat fixable operational problems.

The bigger question may be whether the service architecture itself is flawed in allowing such catastrophic failures to occur in the first place.

Technorati Tags:
,

Scroll to Top