Hello Kris: I still not have posted 08 because I felt from your personal mails that you were OK but you did not confirm on the list. Do you have a reason to hold (like you did not yet fully make up your mind?)
All the best, Pascal From: Pascal Thubert (pthubert) Sent: vendredi 8 mai 2015 17:59 To: 'Kris Pister'; [email protected] Subject: RE: [6tisch] last changes on Archie Hello Kris: Pascal - I'm not on expert on IESG review, so take my comments with a grain of salt. High level: * this feels more like a research proposal or DARPA quarterly status report than an RFC. It's filled with lots of good ideas, and evidence that people are working on solving hard problems, but much of it is *very* preliminary. It is, Kris, and by design I'd say; Archie is a high level view of things and how they bind together. It is also the repository of the work as we progress it, and the design decisions as we make them. * I'm worried that the scope of the document goes way beyond what 6TiSCH was chartered to do. Maybe that's a good thing in managing our path forward with the IESG? I do not think so. Archie is not about 6top. The charter includes the backbone operation, at least for the architecture level. But for the components that actuate the architecture, we can only go to the relevant working groups and suggest additions / modifications as appropriate. * The first section says that we have (using present tense) three different ways to compute routes, four different ways to manage cell schedules, and three different forwarding models, or 36 different combinations. Most of them are still not defined. As an implementer, that really scared me. Will the IESG feel differently? I'm not sure that you'll ever find all of the combinations, some just do not make sense. But these are the tools that we can play with and it' important that they are explained somewhere before we start using them. * What little is specified generally relies on IDs, not RFCs, and even the IDs are often not adopted by their respective working groups yet. Some specifics: * move security from 5.1 and put it in 13 with the rest of security I has that comment from René. Are you looking at Archie 07 - there is a weird thing in the tools that it appears separate? In 07 security is moved down as you suggest, what's left is just the explanation of the boxes that are represented above. We seem to have reached consensus with René, do you wish to reopen? * all of section 6 "6lowpan (and RPL)" seems to be good stuff, but higher layer than 6TiSCH, and with no particular impact on, or from, TSCH. It seems that it should just be removed to a separate draft in another working group. We are specifically chartered to describe how things work, Kris. This includes RPL etc... The problem that we address is that how things work together is described nowhere and we added that to the 6TiSCH lot. * I didn't understand section 7 at all. What are we trying to say here? It's really defining terms and an intro to section 9, 10, 11. I could move that down as section 8 and embed 9, 10, and 11, if that makes more sense? * Section 10 on Forwarding Models seems like an interesting vision, but its so far beyond anything that anyone has implemented (or even discussed? Maybe I missed the threads) that it should be written up in a separate ID. This was discussed many times at 6TiSCH, Kris. And really, one is plain IP, one is the emulation of industrial protocols, and the last (fragments) may never see the light of day, but it is present in drafts and even books so we need at least to describe it at a high level. Again, I don't know the process, so maybe this kind of visionary document is what the IESG is looking for. I'd be much happier with a bottom-up development, e.g. this is what we have now, this is what runs, here's a little bit of extrapolation of where we're headed next. It is an informational document, so it is not normative. It is reporting the status of our work and providing a high level description of how things work together. I do think we are OK there... Thanks for all! Pascal ksjp On 4/29/2015 5:21 PM, Pascal Thubert (pthubert) wrote: Dear all : At the interim call we found that Archie is mostly ready to ship but for 2 things; 1) we need to decide on the IEEE reference 2) some editorial changes would be required to improve readability and make the IESG review process smoother. For 1) the proposal is to use an undated reference since the spec does not reference specific points in IEEE specs. For 2) Kris, as a coauthor, will propose some changes and come back to us in the coming week. If you object to these resolutions, please let us know. Cheers, Pascal _______________________________________________ 6tisch mailing list [email protected]<mailto:[email protected]> https://www.ietf.org/mailman/listinfo/6tisch
_______________________________________________ 6tisch mailing list [email protected] https://www.ietf.org/mailman/listinfo/6tisch
