ping to point-to-point (serial) interface

CARLO CIANFARANI · ‎10-31-2011

Hi,

in the simple Lab

R1--------------------R2

the routers are linked together by serial interfaces (s0/0). I suspected that a ping from R1 for its s0/0 configured ip address is actually physically sent on the wire to R2 and received back by R1.

So, to see this, I've configured same ip address on both i/f (10.0.0.1/30) and turned on debugs

R1#ping 10.0.0.1 rep 1

Type escape sequence to abort.

Sending 1, 100-byte ICMP Echos to 10.0.0.1, timeout is 2 seconds:

!

Success rate is 100 percent (1/1), round-trip min/avg/max = 120/120/120 ms

R1#

R2#debu ip pack det

IP packet debugging is on (detailed)

R2#

*Mar 1 00:34:47.767: IP: tableid=0, s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), routed via RIB

*Mar 1 00:34:47.771: IP: s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), len 100, rcvd 3

*Mar 1 00:34:47.775: ICMP type=8, code=0

*Mar 1 00:34:47.775: IP: tableid=0, s=10.0.0.1 (local), d=10.0.0.1 (Serial0/0), routed via RIB

*Mar 1 00:34:47.779: IP: s=10.0.0.1 (local), d=10.0.0.1 (Serial0/0), len 100, sending

*Mar 1 00:34:47.783: ICMP type=0, code=0

R2#

Others p2p interface type exhibit same behaviour different from, for example, ethernet interfaces

Someone can explain me why these differences between L2 medias ?

Thanks.

Giuseppe Larosa · ‎10-31-2011

Hello Carlo,

your understanding is correct.

p2p L2 media have the behaviour that you have shown in the debug output commands.

ethernet controllers implementation is different and a ping to the IP address of a local LAN based interface is not sent out on the wire:

ethernet is a broadcast and multicast capable L2 media, other media like a FR or ATM network may be multi access / multi point but they are not broadcast and multicast capable by themselves (they can use replication of a frame over different PVCs to emulate it, but it is a different matter).

years ago it was possible to use a PA 4 Eth or a a PA 8 eth to do tests on a C7500 or C7200 by simply disabling the keepalive function ( instead of creating additional loopback interfaces) I did it some times in some functional tests.

things are different for GE or 10GE ports that are not so easy to foul

Hope to help

Giuseppe

CARLO CIANFARANI · ‎10-31-2011

Thanks Giuseppe,

Clearing the ip fast-switching cache on R2 (c3725) and configuring now 2 different subnet ip address for R1 and R2 s0/0 (10.0.0.1 e 10.0.0.2), I see

R2#cle ip cache

R2#sh ip cache

IP routing cache 0 entries, 0 bytes

11 adds, 11 invalidates, 0 refcounts

Minimum invalidation interval 2 seconds, maximum interval 5 seconds,

quiet interval 3 seconds, threshold 0 requests

Invalidation rate 0 in last second, 0 in last 3 seconds

Last full cache invalidation occurred 00:00:03 ago

Prefix/Length Age Interface Next Hop

R2#

R1#ping 10.0.0.1 rep 1

Type escape sequence to abort.

Sending 1, 100-byte ICMP Echos to 10.0.0.1, timeout is 2 seconds:

!

Success rate is 100 percent (1/1), round-trip min/avg/max = 44/44/44 ms

R1#

R2(config-if)#do deb ip pack det

IP packet debugging is on (detailed)

R2(config-if)#

*Mar 1 00:28:23.607: IP: tableid=0, s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), routed via RIB

*Mar 1 00:28:23.611: IP: s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), len 100, redirected

*Mar 1 00:28:23.615: ICMP type=8, code=0 <--------------------------------------------------------------------- ICMP echo rqst

*Mar 1 00:28:23.615: IP: tableid=0, s=10.0.0.2 (local), d=10.0.0.1 (Serial0/0), routed via RIB

*Mar 1 00:28:23.619: IP: s=10.0.0.2 (local), d=10.0.0.1 (Serial0/0), len 56, sending

*Mar 1 00:28:23.623: ICMP type=5, code=1 <--------------------------------------------------------------------- ICMP redirect

*Mar 1 00:28:23.623: IP: s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), g=10.0.0.1, len 100, forward

*Mar 1 00:28:23.631: ICMP type=8, code=0

R2(config-if)#

*Mar 1 00:28:23.683: IP: tableid=0, s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), routed via RIB

*Mar 1 00:28:23.687: IP: s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), len 100, redirected

*Mar 1 00:28:23.691: ICMP type=0, code=0 <--------------------------------------------------------------------- ICMP echo reply

*Mar 1 00:28:23.691: IP: s=10.0.0.1 (Serial0/0), d=10.0.0.1 (Serial0/0), g=10.0.0.1, len 100, forward

*Mar 1 00:28:23.695: ICMP type=0, code=0

R2(config-if)#

R2#sh ip cache

IP routing cache 1 entry, 176 bytes

12 adds, 11 invalidates, 0 refcounts

Minimum invalidation interval 2 seconds, maximum interval 5 seconds,

quiet interval 3 seconds, threshold 0 requests

Invalidation rate 0 in last second, 0 in last 3 seconds

Last full cache invalidation occurred 00:02:23 ago

Prefix/Length Age Interface Next Hop

10.0.0.1/32 00:00:03 Serial0/0 10.0.0.1

R2#

So next ping pkt will be fast-switched and thus no msg will be logged in the debug.....

However It's not clear to me if exist a specific reason for a different implementation between p2p and multiaccess/multipoint media.....