search

lp-wan / datamodel

use of yang to describe SCHC
1 stars 3 forks source link

Terminology Section for AC - things defined in other RFCs #11

Open MarinoMtz opened 1 year ago

MarinoMtz commented 1 year ago

Goal: Align with 8724 and if possible align with architecture draft as it's an ongoing work.

  1. A Context is Set a of Rules [8724]
  2. A Context does not contain additional information [8724].
  3. There is a context or Set of Rules per session [Archi]. 3.1 There can be multiple sessions or instances on a SCHC Core [defined as Network Gateway in archi]. 3.2 A Device typically has only one instance.
  4. A SCHC instance (or session) is a protocol operation between a pair of peers [Archi].
MarinoMtz commented 1 year ago

Do we actually need all of this terms in AC?

LT: Context, with extra info, we have a way of identifying the other end. LT: The Set of Rules is identically in both ends, but the context may not be identical. LT: The Set of rules allow to apply compression and fragmentation AM: The SoR already identify the other end, is the same called context (instanciated or not), there is a Context per device instanciated in the core. IM*: A context is a set of rules intanciated (core or device), and in the body of the document we use only context

LT: Context is a a SoR + meta is a context?

This new vision of context is not the same as in the RFC, since there are some meta data that can change, during the same session (TCP Flow, or changing rules with CORECONF for instance, these are new using SCHC)

AM: SCHC (Static Context.....), if we say that the context may change but is the set of rules that does not change in time. Ana it can be STATIC but it can also change in time, static means that it does not change during a SESSION.

AM: Yes but we are trying to change it during the same session, so ...

minaburo commented 1 year ago

AM: SCHC is a static context because the Rule definition of the Field identifiers is fixed. This is only true when compressing IPv6 and UDP but may differ when other protocols are used. This is the case of CoAP, where Uri-Path and Uri-Query field ids may vary. And so the definition of the FID in a Rule isn't fixed anymore, or at least for those fields.

On Wed, Jun 7, 2023 at 11:41 AM Marino Martinez @.***> wrote:

Do we actually need all of this terms in AC?

LT: Context, with extra info, we have a way of identifying the other end. LT: The Set of Rules is identically in both ends, but the context may not be identical. LT: The Set of rules allow to apply compression and fragmentation AM: The SoR already identify the other end, is the same called context (instanciated or not), there is a Context per device instanciated in the core. IM*: A context is a set of rules intanciated (core or device), and in the body of the document we use only context

LT: Context is a a SoR + meta is a context?

This new vision of context is not the same as in the RFC, since there are some meta data that can change, during the same session (TCP Flow, or changing rules with CORECONF for instance, these are new using SCHC)

AM: SCHC (Static Context.....), if we say that the context may change but is the set of rules that does not change in time. Ana it can be STATIC but it can also change in time, static means that it does not change during a SESSION.

AM: Yes but we are trying to change it during the same session, so ...

— Reply to this email directly, view it on GitHub https://github.com/lp-wan/datamodel/issues/11#issuecomment-1580318244, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEFFDUXKSMRZHWOKBRYKLALXKBEFVANCNFSM6AAAAAAY5PPGXQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>