Namecoin as a Decentralized Alternative to Certificate Authorities for TLS: The Next Generation
From 35C3 Wiki
|Description||How we improved the attack surface, compatibility, and scalability of Namecoin's replacement for the Certificate Authority system|
|Tags||monero, privacy, anonymity, decentralisation, cryptocurrency, blockchain|
|Processing assembly||Assembly:Critical Decentralisation Cluster|
|Language||en - English |
en - English
(Click here to refresh this page.)
|Starts at||2018/12/28 12:00|
|Ends at||2018/12/28 12:45|
|Location||Assembly:Critical Decentralisation Cluster|
Certificate authorities (CA's) pose a serious threat to the TLS ecosystem. Prior proposed solutions (e.g. Convergence, DANE, HPKP, CAA, and CT) simply reshuffle the set of trusted third parties. In contrast, Namecoin solves the underlying problem: if you know a Namecoin domain name, you can find out which TLS certificates are valid for it, with a threat model and codebase nearly identical to the battle-hardened Bitcoin. One year ago at the 34C3 Monero Assembly, I presented a design (with implemented, working code) for accomplishing this in the real world of uncooperative web browsers, with best-in-class attack surface, good compatibility, and good scalability.
But there was still much that could be improved, ranging from ending our reliance on HPKP API's (which are being phased out), to preventing the browser's TLS implementation from leaving your browsing history on the disk, to sandboxing Namecoin's certificate override code so that it can't compromise non-Namecoin traffic even if exploited, to supporting Firefox and Tor Browser (both of which posed unique challenges), to name just a few. This talk will cover a wide variety of improvements we made to attack surface, compatibility, and scalability. Expect to learn lots of interesting little-known trivia about the innards of TLS implementations, which can be used for unexpected purposes in our mission to rid the world of the scourge that is certificate authorities.