Vulnerabilities in ID tech
Vulnerabilities and Weaknesses in Identity Protocols
Convener: Rick Smith
Notes-taker: Rick Smith
Tags: Information Cards, CardSpace
A little bit of the blind following the blind.
User clicks on card selector – transmit to relying party the information.
If it’s a self-issued card, then the client sends it directly.
If it’s a managed card, the data is still sent by the client, but the client sends a token signed by the “manager,” or ID provider.
Cards run in a protected space so that the contents can’t be sniffed by other unprivileged processes.
Four risk areas:
Native code running on client systems, and/or plug-ins on a browser. Attackers can substitute subverted code and intercept personal memorized secrets that secure the cards, or that are used with IDPs to authenticate a managed identity.
Network based attacks – forged transactions or modified transactions used to spoof identity. Most implementations rely on SSL to protect on these. Not all protocols require SSL in all circumstances where it is needed.
Subverted or malicious relying party – can the RP turn around and exploit the user’s identity to masquerade to another RP?
Spoofed IDP – a variant of the network based attack – can an attacker trick the RP into authenticating a user by intercepting IDP transactions and providing a bogus response
TPM modules – there are 300 million machines with TPMs today – we have a way to install secure software and safely manage crypto keys.