Re: draft-narten-ipv6-3177bis-48boundary-02.txt

[Thomas Narten <narten@us.ibm.com>]

Ralph,

> Thomas - is there any way in which you can quantify "excessive"?

Whether "excessive" is an appropriate word to use is a judgement
call. YMMV. But, the word "excessive" is literally the word I hear
many people (especially in the RIR and operational community) use when
talking about this issue. And I'll note, that even when the IETF went
forward with 3177 originally, I think we'd have to say consensus was
"rough". Thus, even within the IETF, some people thought /48 was
excessive.

> E.g., make some suppositions for an ISP:

> * number of subscribers
> * subscribers per aggregation point (CMTS for cable)
> * length of prefix required for each aggregation point
> * length of prefix to be assigned to an ISP based on the above

What 3177bis is talking about is the size of the assignment to a
single end site, e.g., a home, a business, etc. The above are more
about issues internal to an ISP (or a very very large end site), and
thus the above really relate more to the size of the RIR->LIR (ISP)
allocation. I.e., a different issue.

Thomas