Re: [spring] 64-bit locators

[Alexandre Petrescu <[e-mail]>]

Le 18/12/2019 à 18:16, Ron Bonica a écrit :
> Folks,
> 
> I am warming to the idea of fix-length, 64-bit locators. Benefits follow:
> 
>   * There is no use-case for less specific (e.g., /56) locators
>   * It would make the FUNC part of the address appear in a predictable
>     location. This would facilitate ACLs that match on function.

I think it sounds as if one needs a locator.  That locator must have a 
fixed length.

I suggest that locator to be an IP address whose length is 128bit.

Interpreting a part of an IP address as a locator (be that 65bit or 
other length) might lead to confusion between roles of locator and 
identifier.

And, ACL matching is indeed better on fixed length keys, but I think it 
is even better on full addresses than on parts of addresses.

Alex

> 
> While you might save a IPv6 address space with more specific locators, 
> the savings might not be worth the administrative headache.
> 
> In this case, we should probably change the document to reflect 
> implemented behavior.
> 
>                                                                         Ron
> 
> 
> Juniper Business Use Only
> 
> 
> _______________________________________________
> spring mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/spring
>