Re: [TLS] Issue 16: Alert clarifications

[Nelson B Bolyard <nelson at bolyard.com>]

Eric Rescorla wrote:
> http://www3.tools.ietf.org/wg/tls/trac/ticket/16
> 
> Several people have raised the issue of what alerts must be fatal
> and when they should be sent. NIST suggested that all fatal alerts
> MUST be sent. They also suggested that the following alerts be fatal:
> 
> - bad_certificate, 
> - unsupported_certificate
> - certificate_revoked.
> - certificate_expired
> 
> We discussed this in Prague and looking at the minutes, it looks
> to me like the resolution was as follows:
> 
> - All fatal alerts MUST be sent
> - For the above alerts, if you plan to tear down the connection
>   on that basis you MUST make them fatal and send them
> - Add a warning that some implementations tear down the connection
>   for any alert so warning alerts are dangerous. New implementations
>   SHOULD not tear down the connection for warning alerts.

Why not require that a fatal alert be sent any time that the connection
is going to be torn down due to a protocol error of any kind?
That is, any time an error occurs in the TLS protocol which is severe
enough that the connection will be terminated, send a fatal alert first.
This would greatly improve diagnosis of problems.

-- 
Nelson B

_______________________________________________
TLS mailing list
TLS at lists.ietf.org
https://www1.ietf.org/mailman/listinfo/tls