The MPLS WG Archive

[Brian E Carpenter <brian@hursley.ibm.com>]

I'll observe that we went through a similar debate in diffserv and ended
up speaking, at least colloquially, of "edge-to-edge" to identify things
that happen between two edges of a diffserv domain, as opposed to between
two hosts.

  Brian

"David R. Oran" wrote:
> 
> I've observed in the MPLS meeting and on the mailing list that the words
> "end-to-end" are frequently used with respect to various MPLS features and
> mechanisms.
> 
> I would now like to ovserve that this usage is in conflict with the
> conventional use of this term over many years in the Internet communty in
> general, and the IETF in particular.
> 
> This leads to a number of misconceptions and a number of proposals that
> attempt to apply MPLS as an "end-to-end" service or mechanism, which is not
> IMHO appropriate.
> 
> Therefore I would like to plead that the MPLS comunity use the terms
> "single-hop" and "LSP-endpoint-to-endpoint" as opposed to "hop-by-hop" and
> "end-to-end" in drafts and RFCs on things like compression, VoIPoMPLS, etc.
> 
> Putting my lead BVDs on....
> 
> Dave.