[PWE3] Re: Proposed way to move forward for ATM/MPLS in pwe3

[Danny McPherson <danny@tcb.net>]

> 1. A significant point that resulted in a lot of discussion was the need for
> transport efficiency. It seems there were two distinct   perspectives as
> regards efficiency. Some people believe efficiency is an important
> requirement for their network applications. Others argued efficiency need
> not be a determining factor in view of declining bandwidth costs etc.

I think it was more so that efficiency was indeed considerable, 
but associated efficiency complexity must be factored as well.

> 2. There was also some discussion around applications which require the need
> for AAL 5 PDU as well as SDU frame modes, although here the underlying
> issues seemed somewhat unclear. Some people did not see the need for both
> PDU and SDU modes, but others argued that these addressed different
> applications in respect to the transparency of service requirement 
> and therefore both options should be available to service providers.

Hence the [continued] chair/AD requirements for individual mode 
drafts that contain associated applicability statements.   

> 3. The issue of relative implementation complexity was touched on in respect
> to the various options above. Since the differences between the
> encapsulation formats are relatively minor, most people felt the complexity
> would be negligible in relation to the complexity of implementing overall
> interworking functionality required in control and management planes.

I'm not sure I saw consensus either way on this.

> 4. The email exchanges seemed to provide opportunities for some colleagues
> to vent their spleen on ATM in general, with some dark mutterings about
> "other" standard bodies conspiring to subvert the IETF process!   While no
> doubt entertaining, these issues seem somewhat orthogonal to the central
> technical issues we are trying to solve here, so this hardly needs to be
> considered further.

Then disregard it and let's continue with the tasks at hand...

> 5. A significant point regarding the architecture implications of the OAM
> interworking between ATM and MPLS was raised(by Neil Harrison of course),
> but this has so far not been addressed in any detail in pwe3, although some
> discussion has been initiated in the ITU-T on this point. In general I would
> agree with Neil that the requirements for OAM interworking resulting from
> correct application of architecture principles should be factored in our
> solutions, but this needs further discussion. However, at this stage I am
> not sure this will necessarily impact the encap mechanism being considered,
> but may affect the applicability statements.

If you feel it's relative to PWE3 submit I-Ds, we'll continue discussions
from there.

> To conclude, the email discussions indicated that there are various
> applications being considered for ATM/MPLS network interworking which result
> in distinct requirements for each. The various solutions proposed to date
> meet these requirements based on the simple classification identified
> earlier, namely:
> 
>     1) 
> 
>     2) 
> 
>     3) 
> 
> The applicability statement draft for the AAL 5 PDU mode has been submitted
> by Matthew Bocci recently and would meet these requirements.
> 
> One issue that did not elicit much discussion concerns the layering
> guidelines as described in draft bryant, which we were encouraged to follow.
> Correct layering needs to be factored into the encapsulation mechanisms
> being considered, and would impact the SDU mode encap proposal, which
> currently violates ordered layering of protocol control infomation.

An updated version is to be submitted as a WG document.  We can discuss
and revise as necessary.

> There has been extensive and useful technical discussion of many aspects of
> ATM/MPLS network interworking in the past months and I believe we now have
> sufficient material in pwe3 to move forward, assuming people want to. The
> guidelines specifying the applicability statements and layering of protocols
> agreed in Minneapolis provide the means to do this on a reasonable basis.

And we're still awaiting many of them...

> The applications I outline above appear to be the basic set emerging from
> the discussion, so I propose we move forward using them in conjunction with
> the solutions indicated. So far two of the AS drafts have already been
> submitted by Matthew and Ghassem, and I understand Luca is in the process of
> preparing the other two. This should provide us with the necessary documents
> to arrive at a common solution set.
> 
> I propose we adopt this approach as a way to move forward at this stage.  A
> good first step would be to adopt the AS drafts which have been sent as
> working group drafts so that the discussion can converge rapidly.  Danny,
> how do we make this happen ?

As suggested already, the AS notes should accompany the individual modes
with which there associated.  Once this occurs, the individual authors 
should request that the ID be adopted by the WG AND become a WG item.
If there are no _substantial objections, so shall it be!  :-)

-danny






_______________________________________________
pwe3 mailing list
pwe3@ietf.org
https://www1.ietf.org/mailman/listinfo/pwe3