There are many lessons that can be drawn from Heartbleed but one of the most important is that complexity is the enemy of security and general platforms have a tendency to become rather complex.
Which is one of the reasons why I have been rather suspicious of the 'just use DTLS' approach. Building on DTLS means importing all the current and future complexity of TLS. That seems rather a lot of overhead when all we need to do at the DNS layer is to take a message M, encrypt it under a key and plonk a MAC code on the end. I know that the counter argument is that 'TLS is a known quantity, 20 years have gone into development'. But like the argument that 'open source is better because anyone can review' we are seeing that its not necessarily true. Code can get worse over time as well as better. -- Website: http://hallambaker.com/ _______________________________________________ dns-privacy mailing list [email protected] https://www.ietf.org/mailman/listinfo/dns-privacy
