Some time ago the Skype network suffered a 48-hour outage, rather embarrassing for a company offering a service to individuals, an absolute catastrophe for a company hoping to break into the business VoIP market. The one over-riding need every business has from its phone system is that when you go to place a call, it just works. Not most of the time, or 99.9% of the time, every single time you pick up the phone you get a dial tone.
But 99.9% is amazing, I hear you say. Well, have you ever considered just how bad only achieving 99.9% uptime for your phone system is ? Lets look at the maths. There are 31,536,000 seconds in a non-leap year (60 seconds x 60 minutes x 24 hours x 365 days). So 0.1% downtime equates to 31,536 seconds, or 525.6 minutes or just over 8 hours and 45 minutes. That’s a whole working day without your phone ! And you have absolutely no idea when that outage will happen…it might be in the wee hours of a Sunday morning if you’re lucky, or it might be in the middle of the call that was going to close the most important deal you’ve ever made ! Now that 99.9% doesn’t look quite so good, eh?
One of the major plus points to Skype’s architecture was supposed to be the fact that, with a peer-to-peer network, you inoculate yourself against server issues. The bigger the network, the more ’servers’ (supernodes in Skype-speak) you have sitting on people’s desks, tables, laps, wherever. What could possibly happen that would cause a significant percentage of these disparate machines to fail ? Well, now we know at least one answer to that question…windows update ! The outage was caused by a significant number of Skype supernodes (read your PC) automatically re-booting as a result of windows update. (Click here for the word directly from Skype). Not a worm attack, not a flood in a data centre somewhere, nope, a standard update process. And the scary thing is, because this is a direct result of the peer-to-peer architecture (i.e. a design flaw), there’s not much Skype can do about it, despite their protestations to the contrary. Are you going to re-boot your PC a couple of days late, possibly exposing it to the security hole the patch has updated, just because Skype would rather you waited ? Nah, didn’t think so.
Personally, I prefer to have a bit more control over my critical business services.