EIGRP Bandwidth(in DUAL)

elderkaielderkai Member Posts: 279
I have a question regarding the bandwidth value in the DUAL equation. I know that it's essentially the bottleneck bandwidth, but is it the bottleneck of the topology it belongs to or the bottleneck of the specific router's interfaces? I thought it was the EIGRP topology at first, but then I thought about how no EIGRP router knows the entire layout of the topology.

For example, let's say I have three routers with R1 and R2 connected via Fast Ethernet, while R3 is connected to R2 and R1 through a Gigabit link. In R3's calculation, will it put the bottleneck of R1 and R2's connecting interfaces(100,000kbps) or will it put R3's which are both 1,000,000kbps for the bandwidth value? Thanks in advanced. :)


  • MrBrianMrBrian Member Posts: 520
    You're talking about how R1 and R2 would advertise the network in between them to R3? You're right, R3 will use 100,000 for the bandwidth. This is because EIGRP does cumulative delay, but minimum bandwidth. They would both send an EIGRP Update message to R3 with their delay and bandwidth(100,000) settings of their fa interface for that network. So R1 would say, for example, I've got, with a delay of 10, and bandwidth of 100,000.

    R3 will then get that and check the interface settings for the interface on which it received the update. It will then add its delay to the value from R1 and compare the bandwidth in the update to the bandwidth of its gigabit int, and use the lower (worst) one. Then it will plug those values into the equation to calculate the crazy metric number. So yea, R3 would use 100,000 for the bandwidth. And it will probably load balance to R1 and R2 for that network, assuming they have the same values for the network.
    Currently reading: Internet Routing Architectures by Halabi
  • elderkaielderkai Member Posts: 279
    Oh, okay. That makes sense. ^.^
Sign In or Register to comment.