Stay on Top of Enterprise Technology Trends
Get updates impacting your industry from our GigaOm Research Community
Om sat down this afternoon with Facebook Technical Operations VP Jonathan Heiliger to talk about the social network’s infrastructure and started out by asking how the company managed to withstand the heavy user traffic during its username product launch.
Heiliger said that Facebook had originally planned for people to claim their usernames in an auction-style format, and the project was code-named “Hammer.” In the end, the company decided that usernames would be claimed on a first-come-first-serve basis — but the launch’s code name remained. “We decided to keep the ‘Hammer’ code name because it was going to hammer our site,” Heiliger said.
As we all know, the launch went off without a hitch. It was successful because Facebook had extensively tested the launch on the backend of the site, Heiliger said. During the testing, Facebook isolated the username function from all the other services on the site. Heiliger credited this strategy for the rollout going smoothly. Facebook even delayed the release two weeks to ensure it was successful and had extra hardware standing by in case users jammed the site.
Heiliger had strong words for OEMs and system builders during his chat with Om. To compete with sites like Facebook and Google (s goog), Heiliger said, OEMs and system builders need to be more power- and cost-efficient. “You guys just don’t get it,” he said, adding that Facebook has reaped success from investing heavily in its infrastructure.
Video of the chat is here: