Could bugs like Heartbleed pose issues for biometric authentication?
Could a future bug, with similar implications to that of Heartbleed cause major concern to the future use of biometric security? Following the critical Heartbleed vulnerability in OpenSSL, and reading countless articles online (see below for a few) an interesting conundrum came to my mind.
As we now know, there are three requirements to overcome the effects of the Heartbleed bug on any one server or service:
- Patch the affected software on the affected server
- Revoke and re-issue the SSL certificates (essentially the private keys used to encrypt traffic between two points such as the end user’s browser and a bank’s web server for example)
- Change your password for the affected service/application in case it had been compromised
The conundrum focuses on a problem with step 3 and the use of biometric security measures such as fingerprints, retina scans and potentially new vein-scanning technologies. While these technologies are not heavily in use by consumers today they are becoming more commonplace, many users of the new Apple iPhone 5s (myself included) use a fingerprint to speed up unlocking the lock screen and the potential uses are already on the rise as this technology becomes more mainstream.
Taking a high level look at this, from a pure sequence of events (as opposed to analysing how or where the biometric data is stored and/or transferred and how it may or may not be encrypted), I provide the following hypothetical scenario to consider. In 12 months time lets say you can use your fingerprint or a retina scan to get cash out at an ATM, or to identify yourself to your bank and other providers using your smart phone. The technology is in use for a period of time and after a while a bug with similar consequences as Heartbleed happens to be discovered. At that time there may be no clear evidence of whether the bug has been exploited or not, however this actually becomes irrelevant. Taking a worst case scenario, lets say despite the best efforts of the companies, the multiple layers of encryption and all the other security measures that one of the many supporting components of the authentication process has a bug which has, or could potentially cause your biometric details to be exposed, copied or intercepted.
As our primary form of authenticating ourselves today is using a password, we can simply change our passwords which invalidates the potentially compromised user credentials. As I am sure you can now surmise if we were using biometric authentication, we could not simply change our retinas or fingerprints, these stay with us for life. I will admit this is taking the extreme end of a worst case scenario, with any high level security solution, you would expect several layers of protection, but it definitely poses an interesting question of what can be done to invalidate and then re-issue a biometric credential.
Unfortunately I don’t have an answer, I do hope that this might promote some discussion or at least get the idea in the back of a few peoples minds. If anyone has any thoughts or ideas please let me know as I am genuinely curious as to the answer to the riddle. In the mean time, perhaps it is best for us to all strongly consider who we want to hand over our biometric “prints” to… if they are ever compromised you can’t simply change them.
Articles that prompted my thinking: