One hazard of new mass-market consumer technologies is that they make geeks of us all. So it is with Skype, which has just come up with an exhaustive account of its spectacular pre-Christmas crash. But the implications go much further than the technical. With an IPO on the cards, this has not come at a good time.
In a blog post on Wednesday, Lars Rabbe, chief information officer, said a snowball effect was behind the problems, which blocked most Skype users for 24 hours on 22nd-23rd December. Some Skype servers became overloaded; that delayed the response time to users, triggering a bug which caused some versions of Skype’s client software to crash; when those users tried to reconnect to the system, they put extra pressure on supernodes that were already working overtime, bringing the whole system pretty much to a halt.
The direct financial cost, in the form of credits Skype has offered to its paying customers, is relatively small. A spokeswoman told us that users of its “pay as you go” service would get 30 minutes of free service, worth $1. Subscription customers will get one week of free service.
But, coming as new CEO Tony Bates is trying to dress the company up for an IPO, the timing is not good.
To prove that Skype can finally turn its massive user base into a large and profitable business, he needs to convince business users to adopt the service in far larger numbers. Indeed, from what we’ve heard from one person close to the situation, a large part of the reason for pursuing an IPO in the first place is to set Skype apart from the growing mass of VOIP competitors in the eyes of business customers.
To the extent that it hints at possible flaws in the architecture of the system, last week’s crash will not make Mr Bates’ job any easier.
