The single sign-on protocols that allow users to sign in to a range of websites with their Google or Facebook accounts suffer from security flaws that could allow scammers to log in as somebody else, security researchers have reported.
The researchers, from Indiana University Bloomington and Microsoft Research, say they have found a number of serious flaws in OpenID and the single-sign on system used by Facebook, as well as implementations of those systems at several popular websites. Google and PayPal are among the users of OpenID.
[ Learn how to greatly reduce the threat of malicious attacks with InfoWorld's Insider Threat Deep Dive PDF special report. | Stay up to date on the latest security developments with InfoWorld's Security Central newsletter. ]
"The problem here is that the authentication system makes life easier but it makes security management more challenging," said XiaoFeng Wang, one of the authors of the study.
Using a single sign-on login initiates a conversation between the website a user is currently visiting and the provider of the identifying account. The website asks for certain information to be verified, and the account provider responds with a thumbs-up or thumbs-down. But, as with most conversations, there is room for misunderstanding.
In one of the flaws the researchers exposed, for example, not all websites confirmed that a verification coming from OpenID included all of the items the website asked to be confirmed, such as the first name, last name, and email address. The researchers were able to access the request, delete one piece of requested information (the email address, for example) as it went to OpenID and simply re-insert it in the signed okay from OpenID. In this way, even a hacker who didn't control the email address linked to the user's account on the website in question could log in, and potentially make purchases, using that person's account.
Using Facebook's authentication system, researchers were able to persuade third-party websites that they were somebody else and hijack that person's legitimate Facebook account.
The researchers looked only at a few popular websites in the study. These included Sears, Yahoo, the Web-based project management application Smartsheet, FarmVille's Facebook portal and The New York Times website.
All the security flaws they documented have been fixed, the researchers noted. Wang did not know of any cases in which the flaws he and his colleagues documented were exploited by scammers. But, he said, "it's our feeling that there will be a lot of similar problems" on other sites using single sign-on.
The shared login method is growing in popularity, including among e-commerce platforms. A recent study by Forrester Research found that more than 10 percent of would-be buyers abandon online purchases out of reluctance to create a new account.
Forrester Research analyst Andras Cser said the OpenID platform "has long been a pretty weak platform."
The impact of the Facebook authentication flaw, he said, "was bigger than just Facebook being cracked, because a lot of sites use Facebook Connect for logging in. That's a big problem."
Cser predicted that "the fix will come from using two-factor login." He suggested that login requests from a new device or coming from an unusual location could trigger a second set of security questions or require users to retrieve a temporary password by text message.
The study, "Signing Me onto Your Accounts through Facebook and Google: a Traffic-Guided Security Study of Commercially Deployed Single-Sign-On Web Services," is available online (PDF).