Twitter falls victim to a cascading bug

June 23rd, 2012 - 06:23 am ET by J. G.

Affected by intermittent problems on Thursday, Twitter has ruled out a denial of service attack, instead blaming the fault on a cascaded bug.

Twitterers were unable to access their favourite communication and information tool for a few hours on Thursday. Twitter has now apologised for the incident and provided an explanation.

The problem was discovered around 9am PDT. Twitter was inaccessible for all Internet users while mobile users no longer saw new tweets. The situation was corrected around 10:10am before falling over again a half an hour later. Things returned to normal at around 11am.

The UGNazi (Underground Nazi Hacktivist Group) hacking group claimed responsibility for a distributed denial of service attack (DDoS) against the Twitter service for forty minutes against all of their servers around the world.

In the past, UGNazi has claimed responsibility for attacks against government sites in the United States, the 4chan.org anonymous service (to eliminate paedophiles, according to them) and even Google.com (for fun).

Twitter has nevertheless denied an attack, stating that the outage was caused by a cascaded bug although they haven’t provided any additional information. "A cascading bug is a bug with an effect that isn’t confined to a particular software element, but rather its effect cascades into other elements as well".

In summary, "one of the characteristics of such a bug is that it can have a significant impact on all users, worldwide, which was the case today". Twitter adds: "As soon as we discovered it, we took corrective actions, which included rolling back to a previous stable version of Twitter".

Such an outage which lasts hours has a significant impact on a service like Twitter which has 140 million active users. In the beginning, the service was highly criticised for their numerous outages (with the picture of a whale). This is no longer really the case today, with Twitter being reliable and stable over the last six months with uptime of between 99.96% and 99.99%.