Hackers redirected Twitter.com's traffic to a rogue Web site for more than an hour early today by accessing its DNS records using an account assigned to Twitter, the company that manages Twitter's DNS (Domain Name System) servers said today.Twitter initially blamed the early-Friday hour-long blackout of its site on changes made to the company's DNS records, which act like a telephone directory to match the twitter.com domain name with the IP addresses used by its servers.
"Twitter's DNS records were temporarily compromised, but have now been fixed," the company said on its service status page at 2:30 a.m. ET. "We are looking into the underlying cause and will update with more information soon." The status page has not been revised with more information since then.
Twitter uses a New Hampshire firm, Dyn Inc., to manage its DNS records, which match Twitter's domain name (twitter.com, and numerous others) with the IP addresses of its servers.
Today, Dyn denied that its infrastructure had been hacked. Early Friday, Tom Daly, Dyn's chief technology officer, told the Washington Post it appeared someone changed Twitter's DNS records to point visitors to a different IP address using the proper account credentials assigned to Twitter.
"Someone logged in who purported to be a legitimate user of their [DNS] platform account and started making changes," Daly told the Post 's Brian Krebs . "It was not a failing on our systems whatsoever."
Kyle York, Dyn's vice president of marketing, echoed that in an interview with Computerworld . "No unauthenticated e-mail address associated with the account accessed the [Twitter] account," York maintained. "This was not an unauthorized breach of our system."
When asked whether the Twitter account had been used by someone authorized to do so, or if those account credentials had been pilfered by hackers, York declined to answer directly. "You'll have to read between the lines," he said. However, he did point to a tweet on Dyn's own Twitter feed as having the right explanation.
That tweet referenced a story on The Tech Herald , in which reporter Steve Ragan used the clues available, including Dyn's public statements, to theorize that someone compromised a Twitter staffer's e-mail account, presumably via malware that snuck onto the Twitter employee's computer, or through a standard phishing-style identity theft attack.
Once in control of the e-mail account, the hackers then used it to request a password reset for Twitter's account with Dyn, Ragan speculated. "The password reset process is completed, and at this point the person(s) posing as a Twitter staffer gets the reset password via e-mail," Ragan wrote.
That approach makes the most sense, agreed Ray Dickenson, chief technology officer at security vendor Authentium. "That's the most logical explanation," said Dickenson. "If someone obtained administrator credentials for Twitter's account with Dyn, or even if it was inside job, everything worked except the human element."
Dickenson said Dyn's claim that its servers had not been officially hacked is also likely true. "It's very difficult to directly hack a top-tier DNS provider," he said, noting that security at such firms is extremely tight. "You've got to believe that Twitter looked at the options, and made the right choice when it went with Dyn. Twitter's a huge site, and a huge brand."
Also in Dyn's favor, said Dickenson, is the company's contention that only Twitter's DNS records were altered, a fact that York stressed. "The fact that virtually all of Twitter's records were pointing to this defaced site, and that no other [Dyn] customers' records had been altered, corroborates what Dyn's saying."
Twitter has been on shaky security ground for some time. Last August, determined distributed denial-of-service attacks knocked it offline for several hours. Two months before that, a hack of a URL-shortening service redirected millions of Twitter users to an unintended destination.