Digital Consumption


 Charles Darke | 23 January 2007
I'm currently testing a simple fix to prevent phishing attacks on OpenID from malicious relying parties.

The idea is simple: I have a static IP address and so instead of asking for a password, I've modified an IdP to authorise by checking my IP instead of requesting a password. This has the obvious disadvantages of requiring a static IP and limiting you to posting from home.

For posting away from home, an alternative is to install a certificate in the browser and authenticate using this.

It's probably still wise to have passwords as a fallback option but you can be extra careful if you need to enter a password.

Update: I've put up proof-of-concept code here. The code is a hack but it shows that it can be done. 


Comments (4)

IP6 here we come.
  Written by Guest: Don Park at 3:10am, 25 January 2007.
I've updated to include proof of concept code.
  Written by Charles Darke at 10:14pm, 7 February 2007.
This is kind of cool, but what about people who do NOT have a static IP or are logging in from a public computer such as at school, work, or a cybercafe?
  Written by Guest: Andrew at 9:19pm, 28 March 2007.
This is kind of cool, but what about people who do NOT have a static IP or are logging in from a public computer such as at school, work, or a cybercafe?

Again, it's just an idea rather than a recommendation. Obviously, this doesn't work when you're away from home unless you SSH/VNC/RDP into your home computer.

Otherwise, you will have to fall back onto something else e.g. browser certificates (which requires taking your computer (or at least a copy of the certificates) with you.

Ultimately, you'll probably have to fall back to passwords.
  Written by Charles Darke at 1:05am, 29 March 2007.