Bandwidth Command

[jim_berlow]

I have always been under the belief that we should always configure the bandwidth statement under interfaces (especially the logical tunnel interfaces). This idea is founded on the principle that many routing protocols and QoS parameters utilize this information to paint an accurate assessment of the capabilities of a certain interface.

However, I am unable to find any kind of documented best practice stating that. Does anyone know of any document which specifically points out that this is a best practice.

I ask because I have now been told that you should never configure the bandwidth parameter except for the tunnel interface.

Anyone have any documented best practice one way or the other?

Jim

[Giuseppe Larosa]

Hello Jim,

>> I have now been told that you should never configure the bandwidth parameter except for the tunnel interface.

as in any field networking provides space for different thoughts and alternate ways to perform some tasks.

For example, if you use EIGRP on a point-to-multipoint FR you may need to modify the bandwidth command to avoid to saturate the link on hub side (if oversubscribed).

With OSPF also changing the bandwidth to reflect the real speed of a metro ethernet or L3 VPN access link (instead of using ip ospf cost command)

In modular QoS also if you use commands like priority percent, bandwidth percent you need to modify the bandwidth again if you are using a link that is for example an FE or a DS3 but subrate.

Here it is link to command reference 12.4

http://www.cisco.com/en/US/docs/ios/interface/command/reference/ir_a1.html#wp1013415

What is important is that this a logical parameter that cannot influence the real speed of a serial interface.

If you have a good experience in using the command you can keep to use it.

Hope to help

Giuseppe