Many of you will already know that just before Christmas, there was a large scale outage on the Skype voice and video network. The reason given was that their “supernodes” (i.e. customer PCs which are on high bandwidth connections that then get used to coordinate other users) had a software bug that took service out for most of the day.
Many of those who doubt the power of distributed networks or cloud have been saying this marks the beginning of the end of Skype and other similar services.
However, we would argue this merely highlights three key points about services such as Skype and far from marking the end, merely shows what many of us always knew;
1) Skype is fallible, like any other IT or telecoms service.
2) Skype relies upon supernodes, which in a “traditional” VoIP or telephony network would be the core switches. The difference is that Skype uses some of its customers as its switching fabric rather than running and managing it themselves.
3) Any technology platform needs careful rollout of new software, with testing, acceptance and delivery into the live network carefully managed. Skype simply suffered from a bug affecting a large number of core nodes. Being a distributed, Peer-to-peer network did not protect them from the need to write and roll out code carefully.
So what does this mean? Well, if you want to consider who you use to operate IT or telecoms infrastructure, you simply need to understand what the support and service model from your technology partner is. If the voice and video service is not crucial to your business, then the day or so of outage at Skype will have annoyed rather than hurt its customers.
If you rely on the network, it is clear that distributed, Peer-to-peer models may not be appropriate and a fully managed service is probably worth investing in.