Hello Charles
The metric that is calculated for each possible path to a particular destination network using EIGRP is determined using the EIGRP formula as described in this lesson. That formula is based on the bandwidth, load, delay, and reliability of the links along the path. These are represented by K values 1 through 4. (K5 is actually not used as part of the best route calculation, but that’s a different story).
Now the question is, where does EIGRP get the information about these four characteristics? These values are all maintained on a per-interface basis on all the interfaces that are traversed along the path to the intended destination. Take a look at the example output of the show interfaces command below:
Router#show interfaces fastEthernet 0/0
FastEthernet0/0 is up, line protocol is up
Hardware is AmdFE, address is cc02.58a9.0000 (bia cc02.58a9.0000)
Internet address is 192.168.12.1/24
MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,
reliability 255/255, txload 1/255, rxload 1/255
Here you can see the following values:
- Bandwidth, labeled BW is at 100000 Kbit
- Delay, labeled DLY is 100 microseconds
- Reliability is at 255/255
- Load is shown as txload and rxload each at 1/255
Now, the bandwidth and delay are actually values that you can manually change. They’re labels that are assigned particular values by default based on the type of interface. For example:
- a 10Mbps link will have a BW of 10000 Kbit and a DLY of 1000 usec
- a 100Mbps link will have a BW of 100000 Kbit and a DLY of 100 usec
- a 1Gbps link will have a BW of 1000000 Kbit and a DLY of 10 usec
But you can go in and change these values manually. It doesn’t change the actual bandwidth of the interface, but it does affect the metric that is calculated by EIGRP and other routing protocols for traffic that traverses this interface.
Now reliability and load are calculated based on the actual number of errors detected on the interface and on the actual traffic on the link, respectively. These values are updated every 5 minutes. More info on all of these can be found in the following lesson:
So you see, you can affect the metric that EIGRP calculates by changing the bandwidth and delay values of the interfaces along the path to the intended destination. Now, what effect does that have on the variance for load balancing? It doesn’t affect the variance directly, but it will affect the metric measured for a particular path. That can change whether a particular path is considered part of the load-balancing arrangement or not depending upon the value of the variance. Does that make sense?
I hope this has been helpful!
Laz