Re: draft-ietf-ngtrans-ipv4survey-02.txt

[Marc Blanchet <Marc.Blanchet@viagenie.qc.ca>]

-- jeudi, septembre 19, 2002 10:14:55 +0900 Jun-ichiro itojun Hagino 
<itojun@iijlab.net> wrote/a écrit:

> > On Wed, 18 Sep 2002, Marc Blanchet wrote:
> > > -- mercredi, septembre 18, 2002 23:32:17 +0300 Pekka Savola
> > > <pekkas@netcore.fi> wrote/a ?$BqD?(Brit:
> 	this is a slide i will be using tomorrow.  hope it is useful.
>
> itojun
>
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> %page
>
> 32bit IDs
>
> 	A lot of protocols use 32bit IDs
> 	32bit IDs = management headaches, scalability limitation
>
> 	What is the domain of uniqueness?
> 		OSPF - within an AS
> 		BGP - worldwide uniqueness needed
>
> 	How wide does ID need to be?
> 		128bit - global IPv6 address can be used
> 		64bit - EUI64 maybe? (not guaranteed to be unique)
> 		64bit - 32bit AS number + 32bit serial (management headache)
> 		32bit - insufficient (32bit AS numbers)
very good start.
one note about eui-64: for any "server-permanent-like" service, generic bcp 
is to not use the eui-64 as the published address so that in the event of 
hardware change, then the published non-eui64 address is still working.

this means that there is a low chance that eui-64 would be a universal 
solution to identifiers. (not that it could not be used, but it has some 
limitations)

Marc.


------------------------------------------
Marc Blanchet
Viagénie
tel: +1-418-656-9254x225

------------------------------------------
http://www.freenet6.net: IPv6 connectivity
------------------------------------------
http://www.normos.org: IETF(RFC,draft),
 IANA,W3C,... standards.
------------------------------------------