Password use needs an overhaul that is driven not by guesswork but by actually understanding the real damage that can be done when password security is compromised, according to a Microsoft researcher.
While many call for replacing passwords altogether with something else, they may be doing so based on little or no hard evidence, says Cormac Herley, a principal researcher at Microsoft Research.
Keystroke logging, brute force attacks, phishing and session hijacking are all used to get around passwords, but it would be impossible to draw a pie chart of how much each method was used because nobody knows, he says in a paper on the subject. "We don't know the slice sizes — not even approximately," he says.
In addition to finding out, he recommends other steps that could make password use more effective:
- Quantify harm that password compromise causes and differentiate between the worst case and the average case.
- Offer better user support for passwords so password use is more secure.
- Identify when passwords are not enough, and why, so appropriate alternatives can be developed.
- Devise a method for evaluating alternatives objectively.
Herley's premise is that passwords are so entrenched and are useful in so many ways that they're not going away anytime soon. After all, if they were totally ineffective, nobody would use them.
"While the research community is unable to quantify harm, individual companies presumably have estimates of their losses from ongoing threats," Herley says in the paper "A Research Agenda Acknowledging the Persistence of Passwords," coauthored by Paul C. van Oorschot, a professor of computer science at Carleton University. "Their actions currently reveal a preference for password-related losses as opposed to the uncertainty of alternatives."
Passwords have a lot of upsides: they're free, allow access from any machine with a browser, revoking them is simple and it's easy for users who forget them to reset them, that make it hard to dump them altogether. "No single alternative technology is likely to possess the combination of security, usability and economic features that meets all goals in all situations," Herley says.
He also takes the side of end users who are often criticised for creating weak passwords, reusing them and writing them down where they can be compromised.
A set of strong unique passwords to protect different sites and applications creates extra work for users that may be unwarranted, he says. "Without better user-facing support, passwords represent a growing burden of user effort that is better spent elsewhere," he says.
Password support often consists of information on how to choose strong passwords, advice about how to recognise phishing, admonishments to check URLs carefully and to fend off keystroke loggers with antivirus software and updated software patching.
"Thus, they receive, it appears, the advice that is most easily given, rather than the advice that addresses the harms they actually face," Herley writes.