Re: [mpls-tp] [PWE3] Pseudo-wires: uni-directional o bi-directional?

[Greg Mirsky <[e-mail]>]

Dear Venkatesh,
you've said "I really don't understand the point that both SS-PW and MS-PW
are co-routed bidirectional LSPs."
I think that all MPLS-TP documents properly distinguish between PW and LSP.
The co-routedness of the PW is in regard to the fact that all of its end and
transit points (T-PEs and S-PEs) must be aware of relationship between
forward and backward/reverse directions. The co-routed nature of a PW in no
way determines or requires type of tunnel it is mapped to. For example, the
tunnel could be associated bi-directional LSP the PW still will be viewed as
having a co-routed path.

Regards,
Greg

On Mon, Jun 28, 2010 at 12:24 PM, venkatesan mahalingam <
[e-mail]> wrote:

> Hi Sasha and all,
>
> >> I tend to agree with your interpretation: “co-routed” means “co-routed
> in the immediate lower layer”, and hence PWs (both single-segment and
> multi-segment) are co-routed bi-directional LSPs.
>
> Can we always agree that “co-routed” means “co-routed in the immediate
> lower layer”?
>
> I really don't understand the point that both SS-PW and MS-PW are co-routed
> bidirectional LSPs.
>
> I think, PWs use the LSPs as the transport medium to send its packet to one
> PW endpoint and another PW end point. so, LSPs' (server layer - section)
> co-routed/associated bi-directionality will be applicable for client layer
> (PWs) also.
>
> But for LSPs run over datalink, I think, we don't consider the data link's
> co-routed/associated bi-directionality for LSPs as datalink is used between
> nodes.
>
>
> Can you guys please make the things clear for PWs on layer network?
>
> --
> Best Regards,
> Venkatesan Mahalingam.
>
> *From:* [e-mail] [mailto:[e-mail]] *On
> Behalf Of *Greg Mirsky
> *Sent:* Monday, June 28, 2010 10:21 PM
> *To:* [e-mail]
> *Cc:* pwe3; MPLS TP
> *Subject:* Re: [mpls-tp] [PWE3] Pseudo-wires: uni-directional o
> bi-directional?
>
>
>
> Dear Stewart,
> perhaps I'm applying section/segment terminology to PW without sufficient
> explanation of my view. After reading comments by you and Sasha I've looked
> at SS-PW as a bi-directional link or section. If such presentation of SS-PW
> is valid then it's co-routedness, as you've noted, is obvious. I agree that
> SS-PW can be referred as segment as well but then its co-routedness is not
> self-evident. Interestingly, from the view of SS-PW as a bi-directional
> link/section directly follows bi-directional co-routed essence of MS-PW.
>
> Regards,
> Greg
>
> On Mon, Jun 28, 2010 at 12:26 PM, Stewart Bryant <[e-mail]<https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=mailto%3astbryant%40cisco.com>>
> wrote:
>
> Greg
>
> I do not understand your point.
>
> MS-PW S-PEs serve a purpose in breaking the PW into segments. If you do not
> need that decomposition in a network, there is nothing requiring their use.
> However we should not prohibit their use either, as there are scenarios
> whether this decomposition is of benefit.
>
> If you need an existence proof of this point, remember that we initially
> designed SS-PW and needed to create MS-PW to address requirements of network
> operators that has deployed SS-PW.
>
> Stewart
>
>
>
>
>
>
>
> On 28/06/2010 15:01, Greg Mirsky wrote:
>
> Dear Sasha,
> I too find Stewart's explanation very useful for overall MPLS-TP. And in
> this perspective perhaps the SS PW might be also referred as PW section.
>
> Regards,
> Greg
>
> On Mon, Jun 28, 2010 at 8:56 AM, Alexander Vainshtein <
> [e-mail]<https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=mailto%3aAlexander.Vainshtein%40ecitele.com>>
> wrote:
>
> Stewart,
>
> Lots of thanks for a prompt and unambiguous response..
>
> This issues looks like one more fine point of MPLS-TP-ese to me: Does
> “co-routed” mean “co-routed in the immediate lower layer” or “co-routed in
> all the underlying layers down to the duct”?
>
> FWIW, I tend to agree with your interpretation: “co-routed” means
> “co-routed in the immediate lower layer”, and hence PWs (both single-segment
> and multi-segment) are co-routed bi-directional LSPs.
>
>
>
> Regards,
>
>      Sasha
>
>
>
> *From:* [e-mail]<https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=mailto%3apwe3-bounces%40ietf.org>[mailto:
> [e-mail]<https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=mailto%3apwe3-bounces%40ietf.org>]
> *On Behalf Of *Stewart Bryant
> *Sent:* Monday, June 28, 2010 3:46 PM
> *To:* [e-mail]<https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=mailto%3apwe3%40ietf.org>
>
>
> *Subject:* Re: [PWE3] Pseudo-wires: uni-directional o bi-directional?
>
>
>
>
>
> It seems to me that PWs are associated bi-directional LSPs...
>
>
>
> PWs are co-routed bi-directional since it is required that they go through
> the same xPEs in each direction, and the xPEs knows about the association of
> the two directional components.
>
> Whether they run over co-routed bi-directional, or associated
> bi-directional LSP is a deployment/applicability issue.
>
> - Stewart
>
>
>
> _______________________________________________
> pwe3 mailing list
> [e-mail]<https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=mailto%3apwe3%40ietf.org>
> https://www.ietf.org/mailman/listinfo/pwe3<https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=https%3a%2f%2fwww.ietf.org%2fmailman%2flistinfo%2fpwe3>
>
>
>
>
>
> --
>
> For corporate legal information go to:
>
>
>
> http://www.cisco.com/web/about/doing_business/legal/cri/index.html <https://inowa.aricent.com/owa/redir.aspx?C=af2c518a1d214b41962d6a2e2ec260e5&URL=http%3a%2f%2fwww.cisco.com%2fweb%2fabout%2fdoing_business%2flegal%2fcri%2findex.html>
>
>
>
>