Hi, On Sat, Jun 07, 2008 at 04:46:44PM +0300, Joost greene wrote: > Please help me confirm my understanding of how basic routing and ARP works, > assuming the below setup, I'll replay how I think it works hoping for > correction.
Your assumptions are correct, with one small exception:
> SwitchA receives the request (arp who has) for the IP address of Host B and
> it checks its MAC table but none found so it will broadcast the request to
> all ports and changes the Src MAC to that of the switch port that is
> directly connected to e0 on the router.
The ARP *request* is actually sent as ethernet broadcast, so the switch
just broadcasts the packet. (Since the host does not know the target MAC
address, it *needs* to send it as broadcast...)
> *On Router:*
>
> Router de-encapsulates the Layer 2 frame to find the destination IP of HostB
> and looks it up in its routing table, it finds that it's on the same subnet
> as the directly connected interface e1 and so it decides to send it out
> there.
>
> Router knows now where it needs to send this packet and wants to build a
> layer 2 frame for it; we will rewrite the Src MAC to be of e1 before
> sending.
Actually it will usually build a new MAC header, instead of just taking
the ethernet header and "rewrite fields" (the next hop could be a PPP
link, or such) - but effectively, for Ethernet-Ethernet, this is what
happens.
(The IPv4 TTL will be also decremented, and if zero, the packet dropped)
> Src and Dst MAC has been changed along the way but never the Src and Dst IP
> addresses?
Correct.
gert
--
USENET is *not* the non-clickable part of WWW!
//www.muc.de/~gert/
Gert Doering - Munich, Germany [EMAIL PROTECTED]
fax: +49-89-35655025 [EMAIL PROTECTED]
pgpMZUaxoKzZu.pgp
Description: PGP signature
_______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
