Re: rfc2385

[Alex Zinin <zinin@psg.com>]

Friday, April 4, 2003, 2:59:10 AM, Wijnen, Bert (Bert) wrote:
> Inline
>> -----Original Message-----
>> From: Harald Tveit Alvestrand [mailto:harald@alvestrand.no]
>> Sent: vrijdag 4 april 2003 11:36
>> To: Steve Bellovin; iesg@ietf.org
>> Subject: Re: rfc2385
>> 
>> --On torsdag, april 03, 2003 19:45:27 -0500 Steve Bellovin 
>> <smb@research.att.com> wrote:
>> 
>> > Grepping the RFC directory shows that LDP (RFC 3036) 
>> mandates 2385, too.
>> > Does this change what we want to do?  Or is LDP "close enough" to BGP
>> > that the same reasoning will apply?  (3446 also suggests 2385.  This
>> > spreading use is the reason I don't want to promote 2385.)

Yes. The threat model for LDP is even less scary then for BGP,
because all LDP sessions are single-hop.

>> LDP has far less installed base, and its standards track
>> progress has been EOLed by the WG. 

> That is not correct. The CR-LDP has been EOL-ed, but not the base LDP. 

Correct, these are different.

>> So we're unlikely to face the prospect of Draft status for LDP.
>> 
> So I would suspect that base LDP may still want to move to DS
> later, when also MPLS and such want to advance.

This is my understanding too.

> Maybe Alex and I should check with MPLS WG.

Will do.

Alex


>> that said, it increases the reasons why the statement "Don't use TCP-MD5" 
>> should be made in such a way that people checking out its 
>> status will find it.
>> 
> Yep!

> Bert
>>                 Harald
>>