Roland Bless wrote:
> This breaks IMHO a lot of applications, e.g., DynDNS
> registrations of DSL routers
Some DynDNS clients allow to use port 8245 (JFTR, it's
clear that this isn't your point).
> Is there anything newer than RFC 2775 that one could
> give as strong technical advice to abandon that feature
Check out RfC 4084 (BCP 104) "full internet connectivity".
But I doubt that it helps, chapter 3 only says "MUST be
disclosed", they do this.
Maybe you find an appropriate MUST in the HTTP/1.1 RfC.
> I like to raise the awareness of the provider that this
> feature is _technically_ dangerous, though it may make a
> lot of sense for the marketing people...
Sure, it's also an opt-out scheme, just stating what it is
in public comparing it with similar features of say SpamCast
might help to "educate" them, otherwise many Web browsers
will soon offer to do a dummy HTTP GET on startup. <sigh />
Good luck, Frank
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf