When Skype experienced a massive service outage two weeks ago, it sent a kind of panic to its users and the industry watchers. During the outage, the number of Skype users who are online, which usually fluctuates in the number of 5 million to 8 million users, stayed on a flatline at 1 million. For the more than 4 Million users affected by the outage, the service outage caused inconvenience, frustration, and very probably, disruptions to businesses who rely on the popular peer to peer voice over internet application. But beyond these effects to the users, the outage highlights a very important point about the reality of VOIP: as people and businesses become more and more dependent on VOIP technology and services, VOIP reliability will become more and more important.

Two weeks after the outage, the cause is clearer now, having been discussed and analyzed by the blogosphere. The culprit of course lies in Skype’s server software, specifically, on the part of the Skype system that handles user logins. The bug has been sitting in the code and was just waiting to be triggered. And last August 16, Thursday, the trigger happened. Having just received a security update to the operating system, millions of Skype users who are also using Windows almost simultaneously rebooted as a part of the autmated update. This caused a peak in number of users trying to login to the Skype peer to peer network. The added effect of these millions of skype users simultaneously trying to login and the few available node resources (which are important in a peer to peer network) caused a massive outage to Skype’s servers. The outage lasted for almost three days before things were fixed and the root cause of the problem eventually identified.

While it’s clear that the peer to peer nature of the Skype nature was an integral part of the problem that caused the outage (and the Windows udpate causing a reboot on many users the trigger), this is actually of the reality of any distributed software system with a massive user base. And VOIP software fits very well to this profile. So this brings up an important point: if VOIP is indeed the future of voice telecomunnication, the technology must be improved to the point where outages like this would be remote if not impossible. VOIP, before it can become truly mainstream and replace traditional telephone systems must pass the reliability test with flying colors.

Don't miss a post! Subscribe to the RSS feed or by email today!