Re: Last Call: <draft-ietf-tsvwg-iana-ports-09.txt> (Internet Assigned Numbers Authority (IANA) Procedures for the Management of the Service Name and Transport Protocol Port Number Registry) to BCP

[Paul Hoffman <[e-mail]>]

On 1/27/11 8:12 AM, IETF Chair wrote:
> Originally, two ports were assigned for plain and over-TLS, which for HTTP mapped to two different URL schemes: http and https.
>
> Many people thought that this was a waste of a port, and the STARTTLS approach was developed.  You say that it does not work in some cases, and you seem to be suggesting that we go back to the original way.
>
> Maybe it works in some cases and not others.  Can we say which is which?

In a word: no. We have very little operational experience, and where we 
do, it gives conflicting results. Some mail client developers say that 
POP and IMAP STARTTLS works fine, some say that it is unreliable and so 
they just use the alternate ports.

Note that Cullen's example for where it almost certainly would not work 
is for non-stream UDP. Making UDP developers have to come up with a 
stream-like capability to do a STARTTLS-style single port solution 
defeats the purpose of using UDP. The benefit of "we saved another 
port!" over "we forced someone to make UDP more like TCP!" seems like a 
false one.