Hmm, i dunno... But, neighbour discovery all the same, it's a means of passing packets anyway, and if it's going over a /64 with x hundred systems connected to it, and all x hundred hosts are 'discovering' too it might slow things down; essentially back where you starting with arp congestion really, so i don't see any improvement just because ipv6 uses a different flavour of arp with a different name? Fri, 21 Feb 2003 12:22:54 +1100, Geoff Crompton <geoff.crompton@bjhcontrols.com.au> ×ÙÛÅÕÐÏÍÑÎÕÔÙÊ: > On Thu, Feb 20, 2003 at 11:20:09PM +0000, Daniel O'Neill wrote: > > well, ARPs still take up bandwidth, and I'm not sure the scale you're talking about. I would recommend a tiered structure of gateways, each handling, say, 80bits worth of hosts or whichever you prefer, then route the gateways through each other. This also makes it easier for protocols such as SMB and NMB to function correctly, if they ever function ever again. Also, this allows you to create scopes for any incurred broadcasts you wish to send out: > > > > I was under the impression that IPv6 did not rely on ARP packets. That > it used Neighbourhood Discovery instead. Hence it did not rely on > anything below OSI Layer 3 for functionality, apart from the ability to > broadcast messages. Am I incorrect, or where you making ARP synonymous with > ND? > > Geoff Crompton > > > -- > To UNSUBSCRIBE, email to debian-ipv6-request@lists.debian.org > with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org >
Attachment:
pgpQflmBUq_e9.pgp
Description: PGP signature