GitHub has been hammered by a continuous DDoS attack for three days. It's the "largest DDoS attack in github.com's history." The attack is aimed at anti-censorship GreatFire and CN-NYTimes projects, but affected all of GitHub; the traffic is coming from China as attackers are reportedly using China's search engine Baidu for the purpose of "HTTP hijacking."

According to tweeted GitHub status messages, GitHub has been the victim of a Distributed Denial of Service (DDoS) attack since Thursday March 26. 24 hours later, GitHub had "all hands on deck" working to mitigate the continuous attack. After GitHub later deployed "volumetric attack defenses," the attack morphed to include GitHub pages and then "pages and assets." Today, GitHub said it was 71 hours into defending against the attack.

Seven hours later, GitHub said the attackers changed DDoS tactics and it was experiencing outages again.

Four hours later, GitHub said its services were stable although it was continuing to mitigate against the large-scale DDoS attack. Four hours after that, GitHub said the ongoing attack was affecting connectivity to GitHub Gist.

The attack is aimed at GreatFire and CN-NYTimes projects that help Chinese users get around their government's online censorship, but affected all of GitHub. The traffic is coming from China.

According to analysis by Insight Labs, by a Chinese security researcher living outside of China, "My first thought was someone naughty XSSed the page, so I opened developer tools to find the source of the XSS," Anthr@X wrote. "Almost instantly I saw it was keep trying to load these two URLs: github.com/greatefire/ and github.com/cn-nytimes/ every a few seconds."

Baidu is like China's Google and many Chinese website using tracking and advertising code from Baidu. The "HTTP hijacking" was explained as "a certain device at the border of China's inner network and the Internet has hijacked the HTTP connections" and "replaced some JavaScript files from Baidu with malicious ones" that would load GitHub's GreatFire and CN-NYTimes projects "every two seconds."

CN-NYTimes gets around censorship by mirroring NYTimes and GreatFire has long been fighting against Chinese censorship.

Ten days ago, GreatFire said it was "under attack" and asked for help. Last week GreatFire wrote about having "concrete proof from Google that CNNIC (and by extension CAC) is indeed complicit in MITM attacks." That post pointed to Google Security Engineer Adam Langley saying Google was aware of "unauthorized digital certificates for several Google domains." The certificates were issued by CNNIC (China Internet Network Information Center).

While GreatFire was reportedly "delighted that Google, Microsoft and Mozilla have taken steps to blacklist the intermediate certificate used in the attack," the online censorship watchdog called on "Google, Mozilla, Microsoft and Apple to revoke trust for CNNIC immediately in order to protect Chinese user data and user data worldwide."

More:

'Largest DDoS attack' in GitHub's history targets anticensorship projects