Re: [core] Lars Eggert's Discuss on draft-ietf-core-sid-22: (with DISCUSS and COMMENT)

[Michael Richardson <[e-mail]>]

<#secure method=pgpmime mode=sign>

Lars Eggert via Datatracker <[e-mail]> wrote:
    > ### Section 6.3.3, paragraph 3
    > ```
    > The initial entry in this registry is allocated to IANA:

    > +=============+=========+============+===================+==========+
    > | Entry Point | Size    | Allocation | Organization      | URL      |
    > |             |         |            | name              |          |
    > +=============+=========+============+===================+==========+
    > | 0           | 1000000 | Public     | IANA              | iana.org |
    > +-------------+---------+------------+-------------------+----------+
    > ```
    > I would have expected the initial allocation to IANA (and the regions
    > defined within) to be MUCH larger, given the overall size of the
    > namespace.

IANA can ask itself for more space.
There are some advantages to everyone to keep everyone below 2^32.
Although it's not large, since most entries are delta-encoded.

    > ### Section 6.4.2, paragraph 11
    > ```
    > +=============+=========+==========================+
    > | Entry Point | Size    | IANA policy              |
    > +=============+=========+==========================+
    > | 0           | 1,000   | IESG Approval            |
    > +-------------+---------+--------------------------+
    > | 1,000       | 59,000  | RFC Required             |
    > +-------------+---------+--------------------------+
    > | 60,000      | 40,000  | Experimental/Private use |
    > +-------------+---------+--------------------------+
    > | 100,000     | 900,000 | Reserved                 |
    > +-------------+---------+--------------------------+
    > ```
    > These seem very small as well, given the overall size of the namespace.

And noting that 59,000 is < 2^16, it also encodes smaller.
We kept the best parts for ourselves.

A typical YANG module uses one SID per leaf.
Most YANG modules have less than 50 leaves.

We have 270 modules, I read elsewhere, so maybe we'll eat 13,500 out of the
"RFC Required" space in the first "year" (during bootstrap), and then I'll
bet it goes down to about 200 SIDs/year.
Of course, there are some big hairy YANG modules that probably need way more
than 50, but not many of them.

    > ### Section 6.5.3, paragraph 9
    > ```
    > Early Allocations are made with a one-year period, after which they
    > need to be renewed or will expire.
    > ```
    > In practice, that one year is too short and is already creating
    > frequent IESG management items for extension approvals. Given the
    > many more early allocations, this process will require, this will be
    > disruptive for the IESG.

So is the one-year time going to change?

    > ## Notes

    > This review is in the ["IETF Comments" Markdown format][ICMF], You can use the
    > [`ietf-comments` tool][ICT] to automatically convert this review into
    > individual GitHub issues. Review generated by the [`ietf-reviewtool`][IRT].

I should try this again, it didn't work for me last time :-)


--
Michael Richardson <[e-mail]>   . o O ( IPv6 IøT consulting )
           Sandelman Software Works Inc, Ottawa and Worldwide