Password had been with us for a very long time and has shown incredible persistence. Despite countless attempts and near-universal agreement to replace them, passwords are more widely used than ever. Poor security is obviously the main concern of security experts. However, since even strong authentication technologies are vulnerable to certain attacks, more details on exactly what is required of a replacement is essential.
The U.S. government’s 2011 NSTIC initiative, “National Strategy for Trusted Identities in Cyberspace”, summarizes things concisely: “passwords are inconvenient and insecure”. The summary suggests that the implicit goal is “more security, more usability (at reasonable cost)”. There is little to disagree with here; however, it does not point into the direction that would be a suitable replacement. The resources protected by passwords are diverse, from local and corporate accounts, financial accounts with substantial assets, throwaway email accounts, web forum accounts and so on. Clearly, not all type of accounts have the same security needs. Nor do all people have the same security needs; politicians and celebrities in general may require better protection than others need for banking. What should be the starting point for evaluating technologies for the password replacement?
Evaluating the current vulnerabilities for password authentication system is a good starting point. After all, one of the implicit goal for the new authentication method is more security. While usability and cost are important, they usually take a backseat when increased security is required. The end-users and upper management certainly will disagree, but let us just go with the initial assumption and aim for secure authentication.
Password requirements have changed substantially during the years. Long gone are the short alpha and/or numeric only password, at least should be at resources where security is important. Most, if not all systems allow settings password policies that includes complexity, account lockout after x number of attempts and defines expiration as well. Guessing complex and relatively frequently expired passwords is not that productive. It is more of a “my lucky day” type off guess, if successful.
So, what is wrong with the password? It is vulnerable to key-loggers, social engineering, and password cracking.
Arguably, the client devices are the most susceptible for having the account credentials stolen. The source of this issue is the malware-infected devices that had been with us for a long time and will continue in the near future. The bad news? The compromised host or a mobile device enable cyber-criminals to bypass virtually every two-factor authentication system.
Social engineering is manipulating people so they give up the sought after information. The types of information the “social engineer” is seeking can vary, but usually centers on account credentials, financial information, etc. Once the account integrity compromised, the “social engineer”, or designee bypasses virtually any authentication system.
Password cracking requires the password hash that is stored on the device locally, or on the authentication server. Without password hash, none of the password cracking solution would be able to decipher the password. Cyber-criminals utilize various means to obtain access to the password hash, such as exploiting system vulnerabilities, client devices and social engineering. With the compromised authentication server at their disposal, cyber-criminals are capable of bypassing virtually any authentication system.
Are these password vulnerabilities, or the culpability belongs to somewhere else?
The logical answer is that both the client devices and servers are responsible for the password vulnerability. Securing these devices should be the first step in preserving the integrity of the account credentials. Otherwise, the biometric or other types of authentication methods may not provide the desired level of account security. For cyber-criminals, it does not make a difference, if the stolen account credential is password or fingerprint for example. Well, there is a difference. It is easier to replace the password than the fingerprint. Not to mention that while passwords are unlimited, fingerprints for the end-user in question limited to ten.
Based on history, securing the client devices and authentication servers is not likely to take place anytime soon. In which case, replacing password with other authentication methods may provide a seemingly marginal security improvement. The security improvement might turn out to be temporary in nature. At least until the cyber-criminals develop malware that exploits different authentication methods with ease on a wide scale. Keep mind that there is malware available now that capable to exploit two-factor authentication method.
Do we really need to replace the password authentication method now, without addressing the system vulnerabilities first?