The MPLS WG Archive

["W. Mark Townsley" <townsley@cisco.com>]

Yakov Rekhter wrote:
> Mark,
> 
> 
>>Hi Andy,
>>
>>
>>>I'm curious what additional functionality MPLS in L2TPv3 would get you that
>>>GRE or direct IP encapsulation doesn't already provide.
>>
>>The same two points given in the applicability statement of 
>>draft-ietf-mpls-in-ip-or-gre-02.txt for MPLS-in-GRE vs. MPLS-in-IP hold for 
>>MPLS-in-L2TPv3 vs. MPLS-in-IP. L2TPv3 also provides a method for thwarting a 
>>blind packet insertion attack (in the absence of IPsec) via the Cookie in the
>>L2TPv3 header.
> 
> 
> The purpose of draft-ietf-mpls-in-ip-or-gre-02.txt is to define how
> to run MPLS over (a) IP, and (b) over GRE. That is it.
> 
> It is *not* the purpose of the draft to define how to run MPLS over
> all possible sorts of tunneling technologies (with L2TPv3 being
> just *one* of such tunneling technologies).

Yakov,

The purpose of my email was simply to answer a direct question from Andy.

However, more to your point, as long as the GRE or IP draft is not shutting
the door to L2TPv3, I personally don't mind it being defined in a separate
document.

- Mark

> 
> Yakov.
> 
> P.S. One way to support MPLS over L2TPv3 is to use GRE as an
> alternative encapsulation in L2TPv3.
> 
> 
> 
>>- Mark
>>
>>
>>>At 8/29/2003 10:26 AM -0700, David Meyer wrote:
>>>
>>>>One thing that would appear to be useful is to generalize the
>>>>draft just a bit to  mpls-in-ip-encapsulation and include L2TPv3.
>>>>
>>>>Dave
>>
>>
>>
>>
>