Skype's Ghost of Christmas Present

Carl Weinschenk

Carl Weinschenk spoke to Dan York, director of emerging communications technology for Voxeo and author of "The Seven Deadliest Unified Communications Attacks." Skype experienced a major outage before Christmas. The timing couldn't have been worse for the firm, which is nearing an IPO. According to York, Skype uses a different architecture than most VoIP providers. For organizations, the approach has its strengths and weaknesses.

"... the problem [with the Skype outage] was that too much of the peer-to-peer fabric got torn apart and too many supernodes went down."

Dan York
Director of Emerging Communications Technology

Weinschenk: So, what happened with Skype before Christmas?

York: <strong>In the Dec. 22 timeframe Skype had an outage</strong>. You have to step back to see how Skype is set up. It is a peer-to-peer network. There are no servers installed as in a typical system. [Typically there is a] call server, an IP PBX, something in the middle. People connect to the central devices. This is true of consumer systems and enterprise systems. With Skype, there is no central IP PBX, no central call server. It's a P2P cloud. Everyone's clients connect to everyone else and through mechanisms that figure out how to route calls.

Weinschenk: How do they do this?

York: Through what Skype calls "supernodes" that are out on the public Internet. If I am inside my house and I want to talk to my neighbor across the street and can't just yell to him, I have to have someone outside knock on their door and say, "Dan wants to talk to you" and relay messages back and forth between us. That is kind of the role supernodes play. Supernodes are outside of firewalls and outside of NATS. They are not under Skype's control. They use PCs, Macs, Linux. They are Skype clients running on people's systems on the public Internet. What we understand is it could be any PC, any computer that is outside of the firewall. [We understand that they] won't consume more than X amount of your CPU capacity.

Weinschenk: How did the system crash?

York: What happened is that Skype put out Skype 5 for Windows. People upgraded. Skype did an auto update to encourage the upgrade. What they said in their post-mortem is that there was a bug in that software. When it was operating as a supernode it crashed. About half of its supernodes crashed.

Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.

Resource centers

Business Intelligence

Business performance information for strategic and operational decision-making


SOA uses interoperable services grouped around business processes to ease data integration

Data Warehousing

Data warehousing helps companies make sense of their operational data