EIGRP multipoint and point-to-point sub-interfaces

In your book on CCNP Route on page 71 you state: “If you change the bandwidth on the physical interface your sub-interface will not inherit the bandwidth! You need to specify it on each sub-interface yourself.” However, on the following pages in your examples you configure bandwidth on the physical interface and not the sub-interface. It would seem to me that the bandwidth should then be configured on the sub-interfaces on which the PVCs exist, but it’s not. Can you please elaborate on that? Also, in the multipoint example, you only configure the bandwidth on the hub, why would they not be configured on the spokes as well. In the point-to-point example you do configure the bandwidth on the spokes along with the bandwidth-percent command. Please explain this difference in configuration. Thank you!

1 Like

Hello Mike,

I hope you are doing well. I have not read any of Rene’s books but I have been thinking about picking one up more and more. Anyways from what I understand the bandwidth command on an interface does not actually effect how much bandwidth the interface can/will use, this is all dependent on the interface and cable/circuit being used. The bandwidth command is only used by higher level protocols like EIGRP and OSPF for metric calculation. I have put a link to an article discussing this below. I hope this helps!

Scott Weller

I would recommend buying his book, it’s well worth the investment. Anyhow, I understand that the bandwidth command doesn’t actually alter the physical bandwidth, that it’s just a value used by the protocol. What I’m not grasping though is why on page 71 he makes it a point to state if you configure the bandwidth on the physical interface it doesn’t propagate down to sub-interfaces but then in the following examples he configures the physical interface and states you really don’t have to configure the sub-interface, I’m not grasping the why. I mean, I get he was discussing its lack of inheritance, but I don’t get why in those examples we don’t configure it on the sub-interfaces. I’m also not understanding why in the point-to-point configurations he specifies the bandwidth-percent command on the spokes but does not specify that same command on the spokes in the multipoint example. I hope he or someone else enlightens me soon, I feel as though I can’t move on to the next topic until I grasp this particular item.

1 Like

Ahh I see. I don’t think I have help for you there! I will definitely be picking his book up as soon as possible! I hope Rene or Laz replies to this as well, I am very interested myself now.

Hello Mike

To be honest, I don’t know the specific scenario in Rene’s book, however, I may be able to shed a little light on the issue. It is true that by adjusting the bandwidth setting on the physical interface, the bandwidth of the subinterfaces remains unchanged. However, this does not mean that the bandwidth necessarily MUST be changed in such circumstances. In the example in the book, it may be that a change in bandwidth of the subinterfaces is unnecessary.

In any case, can you share the specific scenario so that we can answer more specifically to all of your questions? I look forward to hearing from you!


Yes, I will share the scenario. I don’t have the book with me now though so I’ll have to share it later. Thank you.

1 Like