China's Olympic ticketing system crashes on first day

The Beijing Organizing Committee for the Olympic Games watched unhappily as their online ticketing system for the games crashed shortly after tickets went on sale to the general public. BOCOG's ticketing site received 8 million page views during the first hour of ticket sales on October 30, with an average of 200,000 ticket requests hitting the system every minute. In addition, more than 3.8 million telephone calls flooded the sales hotline as fans tried to book tickets for the games. "Because of the overwhelming volume of page visits, the technical system was unable to perform the tasks well enough, and many applicants were unable to successfully submit their applications," said BOCOG.

After two hours of operating, only 9,000 tickets had been sold and eventually the ticketing system was taken offline for several hours in the afternoon. After IT staff reconfigured the system, it was brought back online at 5 p.m, only to be finally taken offline at 6 p.m., by which a total of 43,000 tickets had been sold (90% of which had been sold online). At a press conference Wednesday, BOCOG officials said the ticketing system had been designed to handle a mere 1 million visits per hour and 150,000 ticket requests per minute. The ticketing system simply could not cope, they said, promising to announce a revamped ticketing plan November 5.

News source: InfoWorld

Report a problem with article
Previous Story

Privacy groups seek 'do not track' Web list

Next Story

Microsoft Releases Vista SP1 RC Preview on MSDN

10 Comments

Commenting is disabled on this article.

Actually it crashed on the first day of Phase 2 of the Olympic ticket ordering. Phase 1 ended about a month ago, and it started earlier this year.

It doesn't really matter what OS that was running (highly doubt it was Microsoft in China), that server would need some massive clustering and redundancy just in order to survive a hit like that.

Additionally, it depends what kind of gateway (if any) they were using to process the payments, as even if their server might be able to handle the requests, the gateway might not (unless they coded with a pipeline in mind for the payments)

Heh, one could argue that they should be able to imagine the number signing up...
But at the same time, dimensioning a system to take 3000+ requests per second is pretty much trying to handle a DDOS attack. :)

Actually, dealing with DDOS attacks of that dimension should be much easier, because that job is about discarding the packets early in the infrastructure, but in this case it's about processing the requests. Ouch...

Also, obviously if a system like this is to crash, it's likely to do so on the first day, because that's probably the day with the highest pressure.