[Detnet] Re: Draft categorization with Taxonomy, for solution selection process

[Jinoo Joung <[e-mail]>]

Hello Janos, thanks for the comment.
I agree to you that the terms "initial" and "final" are somewhat odd here.

They were defined as follows.
The initial time bound is the earliest service completion time allowed for
a packet.
The final time bound is the latest service completion time allowed for a
packet.
(Note that the service is considered as completed when the packet's last
bit leaves a system.)
They are not about the latency in a node, although indirectly related.

Example:
A token bucket regulates a flow.
A packet from the flow is assigned a time, before which time the packet
service is not allowed to be completed.
This time is determined by the flow state, not by the packet arrival time.
Moreover, it is represented with the absolute time, not by the time
difference between the arrival and departure.
So, it is not about the latency.

If we call this time "the minimum time bound", then I think it gives an
impression that this is about the latency.

How about "Earliest service time allowed " and "Latest service time
allowed "?
They seem to be too long but more explicit.
Or any other suggestions?

Best,
Jinoo



On Thu, Jan 23, 2025 at 8:11 AM Janos Farkas <[e-mail]>
wrote:

> Hi Jinoo,
>
>
>
> Thank you very much for making progress with this one!
>
>
>
> Just some minor comments for now:
>
> The terms “Initial time bound” and “Final time bound” look a bit odd to me.
>
> As I understand,
>
> “Initial time bound” stands for the minimum latency/delay, i.e., “not
> earlier than”
>
> “Final time bound” stands for the maximum latency/delay, i.e., “not later
> than”
>
>
>
> It might be useful to use terms that include “min” instead of “initial”
> and “max” instead of “final”.
>
>
>
> Thanks and regards,
>
> Janos
>
>
>
> *From:* Jinoo Joung <[e-mail]>
> *Sent:* Thursday, December 26, 2024 10:33 AM
> *To:* DetNet WG <[e-mail]>; DetNet Chairs <[e-mail]>
> *Subject:* Draft categorization with Taxonomy, for solution selection
> process
>
>
>
> Hello DetNet WG,
>
>
>
> As you recall, a resolution was made that at most one dataplane solution
> per "family" will be selected and processed as a WG document.
>
> The categorization will be based on the Taxonomy draft (
> https://datatracker.ietf.org/doc/draft-ietf-detnet-dataplane-taxonomy/02/
> ).
>
>
>
> The authors of the Taxonomy draft, with the great help from our design
> team leader David, are proposing the family selection result.
>
> Please find attached the slides that explain the background, the process,
> and the result of the family selection.
>
>
>
> Please provide your comments, suggestions, and questions.
>
> Please feel free to make notes on the pdf file attached.
>
>
>
> Best wishes to all of you for happy holidays.
>
> Jinoo
>
>
>