estimates that between 16 million and 17 million Skype users, or about 80 percent of the people who would be on the service right now, can use it. “We are bringing folks back on in a controlled manner,” he says.

Bates priority is to get the service back up and running and to make sure Skype does not lose the trust of its users. To rectify that, he is publicly apologizing to them for the downtime and will offer users some sort of credits, with more details on that coming out later today. (Yes, free calls!) This is the right approach, and reminds me of what Netflix does when its streaming movie service goes down (free movies).

The ultimate cause of the outage is still unclear (or at least Skype isn’t ready to talk about it yet). When I asked if Skype has ruled out a malicious attack, Bates responded: “We haven’t ruled anything out.” Here is what we do know. Many Skype clients on the open Internet act as “supernodes.” These are directories which act like a big peer-to-peer telephone book helping one Skype client find another. These supernodes somehow were overloaded and went down.

Skype had to put up new supernodes to make up for the outage, and did so by redeploying the servers normally used for group video and offline IM features. Those features are down right now. Pulling the servers from other Skype services seems like a short-term solution until Skype can figure out something longer term.

Whatever the ultimate cause, the incident shows up the weak spot in Skype’s network. Take down the supernodes and the whole service goes with them.

0 0

7 7

Authors: Erick Schonfeld

to know more click here