Re: [Raw] New Version Notification for draft-pthubert-raw-architecture-00.txt

[Janos Farkas <[e-mail]>]

Hi Pascal,

The problem statement could be part of the Gap Analysis Document, which is there in the Charter.
The Existing IETF Technologies and Gap Analysis Document could include:
- problem statement
- existing IETF technologies to address the problem
- gaps

Regards,
Janos

-----Original Message-----
From: Pascal Thubert (pthubert) <[e-mail]> 
Sent: Thursday, April 2, 2020 7:22 PM
To: Rute Sofia <[e-mail]>; Pascal Thubert (pthubert) <pthubert=[e-mail]>; Janos Farkas <[e-mail]>
Cc: [e-mail]; [e-mail]
Subject: RE: New Version Notification for draft-pthubert-raw-architecture-00.txt

Hello Rute
> 
> After reading the draft, and being one of the "wireless" as well as "IP people"
> :), IMO:
> 
> - this is not an architecture, so the term framework is possibly better.

I added .framework to the title as suggested in the 01 that I just posted. I expect to keep on working to add more text that's actually architecture.

> - there would be benefits in merging the content of this draft and the 
> OAM requirements draft.

We had a certain organization in the RAW docs and the charter came out differently, so we are reshuffling text between docs, this your impression on the architecture. The architecture deliverable is new with the charter and we are just starting to publish text for it, using text from the problem statement as context since there will not be a problem statement published .

> 
> Relevant to the research towards industrial environments (e.g., 
> industrial IoT) is that determinism needs to be addressed both on the MAC and IP layer..
> Ideally working together.

Great point. 


> 
> The terminology, namely, reliability and availability, are somewhat 
> confusing...perhaps state that reliability is in this draft defined 
> based on two specific indicators, namely, latency (MTBF) and packet 
> loss (MCF).  Similarly to availability, how is a "path" defined here? E2E path? 1 hop?

A path is a very generic term and usually allows A to reach B, so at least it is not a hop between A and B.
In DetNet, the term covers a complex graph with PRE. At 6TiSCH, the decision was to form a term (Track) to cover the complex graph towards B (almost a DODAG but not exactly) on which we apply PAREO functions. 
This doc inherits from 6TISCH for that terminology.

MTBF is not latency but the time between packets that fail to achieve their contract, the contract being for instance in-order delivery within bounded time, to echo DetNet. Or something else, Reliability is abstract to the SLA.

MCF is not packet loss but losses in a row. If you lose exactly every other packet you have a PDR of only 50% but the MCF is 1.  You need both to express that you never lose more than X packets in a row and that this event (of losing >X packets in a row) is really rare.

I put some more text on reliability in -1. Now we need text on architecture as you point out.

Keep safe!

Pascal
> 
> -----Original Message-----
> From: RAW <[e-mail]> On Behalf Of Pascal Thubert 
> (pthubert)
> Sent: 02 April 2020 14:29
> To: Janos Farkas <[e-mail]>
> Cc: [e-mail]; [e-mail]
> Subject: Re: [Raw] New Version Notification for draft-pthubert-raw- 
> architecture-00.txt
> 
> Hello Janos:
> 
> I'll change the title to add /Framework in 01, no worries.
> 
> I have seen a lot of cross participation with DetNet, including you, 
> so I'm reasonably sure that we have that side covered. RAW also 
> attracts wireless experts that did not participate to DetNet. I do not 
> take it as a requirement that what this resulting team produces is a 
> subset of DetNet and is contained in the DetNet Architecture. I 
> actually expect new approaches that are specific to RAW in addition to those inherited from DetNet and possibly extended.
> 
> If that's correct we'll need a new architecture, and we want the 
> wireless experts to help us produce it. I agree that the architecture 
> is the beginning of a solution, which is composed of many components 
> that the architecture outlines. That's the reason why it must come the 
> best informed party, and hopefully RAW is the group that has formed for that at the IETF.
> 
> All in all I believe that the charter makes  a lot of sense.
> 
> Keep safe,
> 
> Pascal
> > -----Original Message-----
> > From: Janos Farkas <[e-mail]>
> > Sent: mercredi 1 avril 2020 22:10
> > To: Pascal Thubert (pthubert) <[e-mail]>
> > Cc: [e-mail]; [e-mail]
> > Subject: RE: New Version Notification for
> > draft-pthubert-raw-architecture- 00.txt
> >
> > Hi Pascal,
> >
> > I like the approach reflected in the first two milestones in the RAW 
> > charter. I found this multiple times a good way of working; namely, 
> > to collect use cases and then collect the requirements out of the 
> > use cases. I think this helps understanding the problem we are about 
> > to solve. Perhaps requirements is close to problem statement. I'd be 
> > glad to see then what can we do with current IETF technology to meet 
> > the requirements / solve the problem and what are the gaps. This may 
> > help seeing clearer what could be useful to include in a framework/architecture.
> > I understand that these work items are not completely sequential but 
> > run with some overlaps.
> >
> >
> > Since you brought up the BoF, the consensus reached on the BoF:
> > https://datatracker.ietf.org/doc/minutes-106-raw/
> > " If do WG for only first 2 work items, informational, do poll again:
> > For/against was maybe 80 vs 20."
> > where the work items were:
> > "1) Produce informational documents describing deterministic 
> > wireless use cases, in continuation to the DetNet Use Cases document
> >  2) Produce informational documents describing the technologies that 
> > the group will cover (e.g., URLLC, TSCH, 802.11ax/be and possibly LDACS)"
> >
> >
> > As you asked feedback about architecture; honestly, I prefer 
> > framework to architecture. Architecture for me is kind of a 
> > solution. I'm most familiar with DetNet, so I'll take it as example. 
> > I think RFC 8655 DetNet Architecture outlines at a high level the 
> > DetNet solution to the problem coming from RFC 8578 DetNet Use Cases 
> > and RFC 8577 DetNet Problem Statement. Of course, RFC
> > 8655 does not provide the solution details, they are provided by 
> > individual drafts like the data plane drafts; but still RFC 8655 is 
> > a solution at
> high level.
> >
> > The RAW charter says: "RAW is not chartered to work on a solution."
> >
> > So, a framework seems to be better to me than architecture in order 
> > to remain in the scope of the charter.
> >
> > As for framework, the DetNet Controller Plane Framework draft comes 
> > to my mind. In the DetNet WG we discussed with regards to this draft 
> > that it is good, but it should not specify solution details.
> >
> > I know, the charter uses "architecture/framework".
> > I'm just expressing an opinion on which one I'd suggest out of the 
> > two options for the reasons explained above.
> >
> > Best regards,
> > Janos
> >
> >
> > -----Original Message-----
> > From: RAW <[e-mail]> On Behalf Of Pascal Thubert
> > (pthubert)
> > Sent: Wednesday, April 1, 2020 3:27 PM
> > To: [e-mail]
> > Cc: [e-mail]
> > Subject: [Raw] FW: New Version Notification for draft-pthubert-raw- 
> > architecture-00.txt
> >
> > Dear all:
> >
> > The collection of documents that we built for the BoF does not match 
> > the list of deliverables that we have now committed with the RAW 
> > charter. In particular, the charter expects a requirement draft and 
> > an architecture draft but does not require a problem statement.
> >
> > Georgios and I picked and restructured text from the PAREO 
> > requirements and the PS draft, and started an architecture document. 
> > I just pushed the first version to hint the group of where we wish 
> > to go and
> solicit early feedback.
> > Please expect revisions soon to add more architecture text.
> >
> > Comments welcome!
> >
> > Keep safe...
> >
> > Pascal
> >
> >
> > -----Original Message-----
> > From: [e-mail] <[e-mail]>
> > Sent: mercredi 1 avril 2020 15:14
> > To: Georgios Z. Papadopoulos
> > <[e-mail]>;
> > Georgios Papadopoulos <[e-mail]>;
> > Pascal Thubert (pthubert) <[e-mail]>
> > Subject: New Version Notification for 
> > draft-pthubert-raw-architecture-00.txt
> >
> >
> > A new version of I-D, draft-pthubert-raw-architecture-00.txt
> > has been successfully submitted by Pascal Thubert and posted to the 
> > IETF repository.
> >
> > Name:		draft-pthubert-raw-architecture
> > Revision:	00
> > Title:		Reliable and Available Wireless Architecture
> > Document date:	2020-04-01
> > Group:		Individual Submission
> > Pages:		16
> > URL:            https://www.ietf.org/internet-drafts/draft-pthubert-raw-
> > architecture-00.txt
> > Status:         https://datatracker.ietf.org/doc/draft-pthubert-raw-
> architecture/
> > Htmlized:       https://tools.ietf.org/html/draft-pthubert-raw-architecture-00
> > Htmlized:       https://datatracker.ietf.org/doc/html/draft-pthubert-raw-
> > architecture
> >
> >
> > Abstract:
> >    Due to uncontrolled interferences, including the self-induced
> >    multipath fading, deterministic networking can only be approached on
> >    wireless links.  The radio conditions may change -way- faster than a
> >    centralized routing can adapt and reprogram, in particular when the
> >    controller is distant and connectivity is slow and limited.  RAW
> >    separates the routing time scale at which a complex path is
> >    recomputed from the forwarding time scale at which the forwarding
> >    decision is taken for an individual packet.  RAW operates at the
> >    forwarding time scale.  The RAW problem is to decide, within the
> >    redundant solutions that are proposed by the routing, which will be
> >    used for each individual packet to provide a DetNet service while
> >    minimizing the waste of resources.
> >
> >
> >
> >
> > Please note that it may take a couple of minutes from the time of 
> > submission until the htmlized version and diff are available at tools.ietf.org.
> >
> > The IETF Secretariat
> >
> >
> > --
> > RAW mailing list
> > [e-mail]
> > https://www.ietf.org/mailman/listinfo/raw
> 
> --
> RAW mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/raw