I knew I forgot something:

4. I think it would be appropriate for the title of this document to be something that relates it to RFC 8314, for example:

Updates to Use of Transport Layer Security (TLS) for Email Submission and Access

The current title reads as if this document were the main body of recommendations.

Keith

On 11/27/18 9:35 PM, Keith Moore wrote:
1. I especially appreciate this document being expressed as deltas to RFC 8314, as I believe this makes it much easier for a reader (especially an implementor or mail service operator) to understand exactly what's being changed.

2.  For existing accounts that have already been established between an MUA and a mail service, should updated MUAs that follow these revised recommendations continue to regard TLS 1.1 as providing the minimum confidentiality level?    Or should such MUAs now warn users that these services no longer provide adequate confidentiality?   Or are there specific server profiles (e.g. certain ciphersuites) of TLS 1.1 that still provide adequate confidentiality which can be detected by the client?

IMO there is a delicate compromise to be made between:

- encouraging service operators to upgrade their security,

- not encouraging users to worry needlessly about security risks that don't actually exist, and especially,

- not conditioning users to ignore security warnings because they're too frequently present.

So when an updated MUA uses an existing account configuration to make a connection with a server that supports TLS 1.1 at most, IMO it should still regard that connection as providing minimum confidentiality as long as there's no reason to believe that the connection is insecure.  (Updated MUAs will still flag newly established connection to that server as not having minimum confidentiality, so operators still have some incentive to .)

However if the TLS 1.1 connection (for example) is using a ciphersuite now known to be insecure, an updated MUA should begin to flag those connections as no longer having minimum confidentiality.

(The theory that I have is that if security warnings to users are consistently accurate and provide a warning of severity that is proportionate to the threat, users will learn to treat them as credible.)

Keith


_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta

_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta

Reply via email to