Example of a Man in the Middle attack on One Time Password systems

Here's a great example of an attack focused on Citibank a few months ago, where a Man in the Middle exploit was used to defeat a One Time Password implementation of 2-factor authentication.

http://blog.washingtonpost.com/securityfix/2006/07/citibank_phish_spoofs_2factor_1.html

If it's just data that you type in, it can be stolen, spoofed, or passed through, leaving you no better off than before.

Paypal's getting the message

Paypal recently rolled out a 2-factor authentication option:
https://www.paypal.com/securitykey

The trouble here - and with any One Time Password (OTP) device - is that these systems are quite vulnerable to Man in the Middle (MITM) attacks. If a thief can get in between your browser and the server, they can pass through your OTP along with your other credentials without your knowledge, and hijack your session at that point, once the server believes a valid session has been established.

What's really needed here is a second channel of authentication and security, relying on a physical hardware device (Something You Have) that can only be checked dynamically by the software system itself, and that is completely independent of the actual systems you're trying to access (i.e. banking website).

Wikipedia entry for our topic

The Wikipedia entry for 2-factor authentication is not bad. We hope to participate in enhancing it further.

For reference, here's the URL:
http://en.wikipedia.org/wiki/Two_Factor_Authentication

Online Banking

Similar to the previous post regarding online commerce, every time you log into your bank you're conducting a 1-factor authenticated transaction. Sure, Bank of America has "SiteKey" which asks you for additional knowledge-based credentials (Something You Know) but other than a clever image-oriented system to help you avoid entering those credentials into a fake version of their website, this system, and systems like it, don't help protect you. They help prevent your credentials from being stolen, sure, but as we saw in the Dateline piece, those credentials are fairly easy to obtain or guess - and once obtained, you're in big trouble.

What's needed is a true 2-factor authentication system to validate your identity before you log into the bank website.

The Financial Services Technology Consortium issued recommendations on this subject, available at the following URL:
http://www.fstc.org/projects/docs/Recommendations_and_Requirements_for_BMA_v1.0.pdf

We wholeheartedly agree with this document, especially their statement that “no authentication system should rely solely on passwords or other knowledge-based queries or shared secrets. If a password should be compromised, it must not be feasible for an impostor to defeat authentication with just knowledge of a password and an associated claim—e.g., userid, account number, name, SSN.”

Online Commerce

http://www.msnbc.msn.com/id/17822386/

This Dateline piece makes it abundantly clear just how easily a thief can obtain credentials - like a credit card number, password, user ID, etc - and wreak havoc on a victim in a matter of minutes. Some scams are more elaborate than others, but the bottom line is always the same - if you can type in a few pieces of information, the computer on the other end of the line doesn't know the difference between you and the real owner of those credentials.

That's what 2-factor authentication is for. If the merchants in this story required a physical authentication device (Something You Have) in addition to the knowledge-based authentication (Something You Know) the thieves would have no way to use the data they've collected to do any damage.