Ethereal-users: RE: [Ethereal-users] unsuccessful ARP repetition and duplicate IP address
> 1. I tried the following: telnet 10.0.1.33 (which is not existent),
> 1st time I got 3 ARP packets before telnet gave up,
> 2nd and 3rd time I got 6 ARP packets. My question is:
> What is the mechanism in the ARP that decides when to quit
> after several unsuccessful ARP request?
It really isn't the ARP mechanism that is resending, it
is telnet. Take a look at RFC 826
http://ietf.org/rfc/rfc0826.txt?number=826
or W. Richard Stevens TCP/IP Illustrated Vol 1.
> 2. I set 2 PCs' IP address to 10.0.1.22 (PC1 and PC2),
> so there is duplicate IP address,
So you are already in a state of sin...
> then I used the 3rd PC(PC 3) to telnet 10.1.0.22,
> PC1 first sent back a ARP reply to PC3 with its
> MAC address, then PC3 sent a SYN tcp packet to PC1...
> [the] ARP reply from PC2 was just ignored by PC3.
> ...I expected that ARP reply from PC2 would update PC3's
> ARP cache with PC2's MAC...
> Robert
Two points:
1) Any IP packets from PC 1 could be used to refresh
the current working ARP entry
2) The implementation does not need to replace
an ARP entry until it times out.
One of the nice things about Stevens' book is that
he looks at situations much like this and sees what
really happens. The different implementations can
do very different things and still be within spec.
Or not. Sometimes you will find stuff that does not
match the spec, but still works mostly.
- jeff parker
- axiowave networks
