[Raw] Some comments on draft-pthubert-raw-architecture-05

[Lou Berger <[e-mail]>]

Hi Pascal/Authors,

I have a some comments on the current draft - I mentioned these at the 
last WG session but have been remiss in sending them to the list, my 
apologies!

I really have two top level comments (detailed comments can wait until 
these are sorted):

1) The document could use some further alignment with the WG 
charter/direction.  We talked last year [1] about the document having 
some pre-WG text and it would be aligned with WG scope/charter,  I think 
it has improved, but still as work to go.

An example of this is the Abstract.  That's pretty easy to fix, e.g., 
the following, which is taken from the Charter, could be added ad as the 
abstract and the first paragraph of the Intro section:

     Reliable and Available Wireless (RAW) provides for high
     reliability and availability for IP connectivity over a
     wireless medium. The wireless medium presents significant
     challenges to achieve deterministic properties such as low
     packet error rate, bounded consecutive losses, and bounded
     latency. This document defines the RAW Architecture.  It
     builds on the DetNet Architecture and discusses specific
     challenges and technology considerations needed to deliver
     DetNet service utilizing scheduled wireless segments and
     other media, e.g., frequency/time-sharing physical media
     resources with stochastic traffic.

But really this comment should be applied generally.

2) A somewhat related comment is that as an architecture document, the 
document cover the full scope of the RAW problem space and leave room 
for all types of wireless networks and integration/optimization 
approaches.  For example, as I understand it, RAW's scope covers 
technologies where the IETF has zero control over the technology and can 
only operate across a  'user' network interface, e.g., one standardized 
by the IEEE, as well as technologies where the IETF has significant 
freedom to define a customized and an IP-integrated control solution, 
e.g., 6TiSCH.

As I read it, parts of the document are very focused on the latter, 
i.e., certain types of wireless networks and a specific integrated 
optimization approach, without allowing for the former.  To me it seems 
you really have two (valid) documents in one here.  The first is a 
foundation for the RAW architecture document, the second is the basis 
for a framework document for RAW networks whose radio layer and the IP 
layer control/management is fully integrated.  If you split this work 
into two documents, I think we can have the Architectural foundation for 
multiple solutions frameworks, where your integrated framework is one 
and another can be based on the cooperative layered architecture like 
the one I mentioned in [1] (e.g., as discussed RFC5623, RFC7926 or many 
other teas documents, and could leverage DLEP).

What do you think? If you agree to the split, I'm happy to help with the 
architecture document, although I also don't want to be the one to hold 
up progress.

Thank you!

Lou

[1] https://mailarchive.ietf.org/arch/msg/raw/TYq4ZL3RZXRkUedOVCdivNGYKOs/