LinkedIn breach .. the multiplier effect
Last week's high profile theft of password for the business networking site LinkedIn has been a topic of many news articles and blogs in the computing and security area. There was a particularly interesting discussion of the issue of developers not understanding the security implications of using fast hashing algorithms which are intended for high-speed, on the wire, integrity checking such as SHA, for secure storage of passwords. The argument here is that password hashes should actually use more computationally demanding (i.e. slower) algorithms. This would increase the cost to the attacker in conducting a brute force attack to identify the correct plain text to match the encrypted passwords in the stolen file(s).
The immediate countermeasure that was publicised through media and through social network 'word of mouth' was that people should change their LinkedIn password. Of course this created the opportunity for phishing attackers (unrelated to those who stole the passwords in the first place) to try and lure users to their own facade sites to try and get the passwords from users. In the midst of all this, LinkedIn also sent out an e-mail to users instructing them to change their password. According to Cloudmark, over 4% of the users who received this legitimate e-mail thought that it was a potential phishing attack and ignored the advice to change their password (thus thinking they had avoided the phishing attack but remaining vulnerable to the attackers who stole the passwords in the first place). Cloudmark make the point that this is partly due to the fact the LinkedIn have a track record of generating lots of e-mail notifications and therefore users get into the habit of ignoring these e-mails. However, it is not necessarily true that just because the user ignored the e-mail (i.e. marked it as spam/phishing), that they also didn't go ahead and visit the LinkedIn site directly and change their password anyway.
The one piece of advice I haven't seen in all this coverage is a reminder to people to think about their other service subscriptions where they might have used the same credentials as those used for LinkedIn. Given the plethora of online services that people use nowadays, and the difficulty of remembering different passwords for each of them, I wouldn't be surprised if many people use the same username/password combinations across many of them. In these cases, users need to change their login password on those services too. Otherwise they might experience a multiplier effect of attacks on their other online accounts as a result of the LinkedIn breach.