LinkedIn said it would reduce the persistence of cookies it uses to identify users of the business-focused social networking site following the discovery of security issues with the site that create a possible means for fraudsters to hijack profiles.
Security researcher Rishi Narang discovered that LinkedIn session cookies are transmitted over an unsecured HTTP connection even in cases where users follow the option of signing in over a secure (SSL) connection. These cookies remain active for up to a year. Hackers who captured these cookies, perhaps using a tool such as Firesheep to sniff out cookies transmitted over open Wi-Fi connections, would be able to obtain unauthorised access to other users' accounts.
The LEO_AUTH_TOKEN cookie grants access to an associated account irrespective of whether or not users are logged in at the time, Narang warns. These cookies work for up to a year or until a user changes their password and logs in using this new password, generating a fresh authentication token. LinkedIn boasts more than 100 million registered users, a factor that inevitably makes it of interest to miscreants.
In response to the research, LinkedIn reduced the persistence of the authentication cookie from a year to three months. In addition, the business-focused social network is extending plans to support SSL across its site – not just during logins, as explained in a statement below.
Whether you are on LinkedIn or any other site, it’s always a good idea to choose trusted and encrypted Wi-fi networks or VPNs whenever possible. If one isn't available, we already support SSL for logins and other sensitive web pages.
Now, we are accelerating our existing plans to extend that SSL support across the entire site on an opt-in basis. And, we are going to reduce the lifespan of the cookies in question from 12 months to 90 days.
LinkedIn takes the privacy and security of our members seriously, while also looking to deliver a great site experience, and we believe these two changes will allow us to strike that balance.
More details of Narang's research, including sample codes and a two-part video of the exploit in action, can be found in a blog post here. ®