Re: [TLS] prohibit <1.2 support on 1.3+ servers (but allow clients)

[Dave Garrett <[e-mail]>]

On Thursday, May 21, 2015 06:59:17 pm Yoav Nir wrote:
> Server implementers will ignore this even if we make it a MUST.

So be it. As said before, there are no IETF police. Responsible server implementations will follow the spec and the ones who want to trade security for compatibility will not. Certainly not ideal, but it'd be a big improvement over everyone ignoring things until they explode. At bare minimum, when the inevitable new vuln is published, the organizations running compliant implementations will be able to show they took things seriously and everyone else did not. I'd at least like to write down the fact that everyone should be aware that it should be obvious that continuing to use an obsolete security protocol for over a decade and a half is not wise. It would help upgrade security more than ignoring the issue.

On Thursday, May 21, 2015 07:09:42 pm Yoav Nir wrote:
> Then point at the BCP. Only problem is that it doesn’t say what you want it to say (it allows TLS 1.0 when the client does not support anything else)

In my first post I said that I don't propose changing anything for existing deployments. The BCP is currently a "SHOULD NOT" for current TLS. I'm proposing a "MUST NOT" for TLS 1.3+ implementations only.

The BCP is also not directly part of the TLS spec. What I'm suggesting is considering 1.0/1.1 to be not legitimate forever, not just not "best".


Dave