effects of configuring the bandwidth statement on eth int

mgottfried · ‎04-24-2009

I'm interested in learning exactly what the effects of configuring bandwidth on an ethernet interace are. I know that it it will effect routing metrices and utilization percentages on NMS systems. But, does it actually effect the output of the show interface command? For example, does it effect the rx and tx numbers shown. IE: the x/255 numbers.

John Blakley · ‎04-24-2009

The only thing I believe it will affect is going to be the BW statement under show interface. The bandwidth statement is mainly used for calculations.

HTH,

John

HTH, John *** Please rate all useful posts ***

lamav · ‎04-24-2009

No, it does not.

Switch_1#sh int fa0/24

FastEthernet0/24 is up, line protocol is up (connected)

Hardware is Fast Ethernet, address is 000e.8364.6d80 (bia 000e.8364.6d80)

Internet address is 50.50.50.1/30

MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,

reliability 255/255, txload 1/255, rxload 1/255

Encapsulation ARPA, loopback not set

Keepalive set (10 sec)

Full-duplex, 100Mb/s, media type is 10/100BaseTX

input flow-control is off, output flow-control is unsupported

ARP type: ARPA, ARP Timeout 04:00:00

Last input 00:00:06, output 00:00:08, output hang never

Last clearing of "show interface" counters 00:05:37

Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

Queueing strategy: fifo

Output queue: 0/40 (size/max)

5 minute input rate 0 bits/sec, 0 packets/sec

5 minute output rate 0 bits/sec, 0 packets/sec

37 packets input, 4762 bytes, 0 no buffer

Received 28 broadcasts (0 IP multicasts)

0 runts, 0 giants, 0 throttles

0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

0 watchdog, 24 multicast, 0 pause input

0 input packets with dribble condition detected

34 packets output, 4066 bytes, 0 underruns

0 output errors, 0 collisions, 0 interface resets

0 babbles, 0 late collision, 0 deferred

0 lost carrier, 0 no carrier, 0 PAUSE output

0 output buffer failures, 0 output buffers swapped out

Switch_1#

Switch_1#conf t

Enter configuration commands, one per line. End with CNTL/Z.

Switch_1(config)#int fa0/24

Switch_1(config-if)#bandwidth 1

Switch_1(config-if)#^Z

Switch_1#

Switch_1#sh int fa0/24

00:40:33: %SYS-5-CONFIG_I: Configured from console by console

FastEthernet0/24 is up, line protocol is up (connected)

Hardware is Fast Ethernet, address is 000e.8364.6d80 (bia 000e.8364.6d80)

Internet address is 50.50.50.1/30

MTU 1500 bytes, BW 1 Kbit , DLY 100 usec,

reliability 255/255, txload 1/255, rxload 1/255

Encapsulation ARPA, loopback not set

Keepalive set (10 sec)

Full-duplex, 100Mb/s, media type is 10/100BaseTX

input flow-control is off, output flow-control is unsupported

ARP type: ARPA, ARP Timeout 04:00:00

Last input 00:00:00, output 00:00:01, output hang never

Last clearing of "show interface" counters 00:05:58

Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

Queueing strategy: fifo

Output queue: 0/40 (size/max)

5 minute input rate 0 bits/sec, 0 packets/sec

5 minute output rate 0 bits/sec, 0 packets/sec

42 packets input, 5168 bytes, 0 no buffer

Received 31 broadcasts (0 IP multicasts)

0 runts, 0 giants, 0 throttles

0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

0 watchdog, 27 multicast, 0 pause input

0 input packets with dribble condition detected

40 packets output, 4602 bytes, 0 underruns

0 output errors, 0 collisions, 0 interface resets

0 babbles, 0 late collision, 0 deferred

0 lost carrier, 0 no carrier, 0 PAUSE output

0 output buffer failures, 0 output buffers swapped out

Switch_1#

I changed the bandwidth statement from 100 M to 1 K and it did not effect the load at all.

HTH

Victor

mgottfried · ‎04-24-2009

I got in touch with a CCIE I know after posting this and he told me that the bandwidtch statement will affect the routing protocols, QOS, and SNMP responses, but won't affect the sh int output.

lamav · ‎04-24-2009

Um, I just told you that. :-)

Jon Marshall · ‎04-24-2009

Yep you did indeed so a rating is in order :-)

lamav · ‎04-24-2009

Shucks, Jon. After this week, I really needed that....lol

Joseph W. Doherty · ‎04-24-2009

"I changed the bandwidth statement from 100 M to 1 K and it did not effect the load at all. "

Victor, yes your stats show that, but considering the input/output rates and in/out packet counts, wouldn't expect the load indications to change much if at all. I though load stats were based on bandwidth. Could you change the load interval down to 30 sec and push 1 Kbps for 30 seconds and see if stats still stay at 1/255?

lamav · ‎04-24-2009

Joseph:

You make a good point and I had thought about that when I ran my test. I dont have a bit generator and I did that test on the fly, but that is also why I lowered the bandwidth statement so much - 1 kbps. Even with the low throughput, though, were the load effected by the bandwidth statement (which we know for a fact that it is not anyway) it would have increased to a little bit more than a 1/255 ratio.

Joseph W. Doherty · ‎04-24-2009

Victor, I had considered what you suggest before I wrote, i.e. a 1 Kbps would show something compared to 100 Mbps, but I think the 1/255 load might also reflect the load interval. If it does (and if it also does reflect the bandwidth statement), the combinations of low number of packets delta in 5, out 6, time delta 19 s, plus the time from the config change, and plus the 5 minute load interval might be why it might not have changed.

Unfortunately, I don't have a box to try against, so I was hoping you could look again with a lower time interval and with more traffic.

lamav · ‎04-24-2009

OK, Joseph. I'll try it tomorrow and post the result.

[EDIT] By the way, I know for a fact that the bandwidth statement on the interface has nothing to do with the actual traffic load on the port and the port's ability to process it and guage it. These are functions of the interface hardware, clocking, and drivers. [EDIT]

Goodnight.

Edison Ortiz · ‎04-25-2009

But, does it actually effect the output of the show interface command? For example, does it effect the rx and tx numbers shown. IE: the x/255 numbers.

It does

Rack1R1#sh int f0/0 | i tx|rate|BW

MTU 1500 bytes, BW 1 Kbit, DLY 100 usec,

reliability 255/255, txload 242/255, rxload 242/255

Queueing strategy: fifo

30 second input rate 14000 bits/sec, 0 packets/sec

30 second output rate 14000 bits/sec, 0 packets/sec

!

Rack1R1#conf t

Enter configuration commands, one per line. End with CNTL/Z.

Rack1R1(config)#int f0/0

Rack1R1(config-if)#no band 1

Rack1R1(config-if)#end

!

Rack1R1#sh int f0/0 | i tx|rate|BW

MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,

reliability 255/255, txload 1/255, rxload 1/255

Queueing strategy: fifo

30 second input rate 14000 bits/sec, 0 packets/sec

30 second output rate 14000 bits/sec, 0 packets/sec

HTH,

__

Edison.

lamav · ‎04-25-2009

I wonder then if it has to do with the platform and the version of IOS. This is a first for me. I actually tested this several years ago in a production environment that had a considerable amount of traffic flowing through the interface and I did not see this effect.

Victor