eigrp's factor of 256...

it is often written that eigrp is an improvement over igrp by a factor of 256...

hence, multiplying by 256... russ white:


remember that EIGRP uses the IGRP metric as its reference point and that IGRP uses a 24-bit number—2²⁴ is its maximum metric—to represent the distance to a destination. EIGRP’s designers wanted to use 32 bits to add granularity to the metric. The simplest way to use the 32 bits was to shift the existing IGRP metric left by 8 bits, which is equivalent to multiplying it by 256.


mtu, load and reliability, when configured cannot trigger an update and are therefore not useful in the calculation... on the contrary, retooling bandwidth and/or delay will trigger an update

you've seen what follows many different ways... this is arguably best in show...
take it away, russ:

The first thing to do when figuring out which path A will choose is to see what metric A will calculate for each path.

• The path through B will be ((10⁷ / 10,000) + 100 + 100 + 100)256, which is 332,800.

• The path through D will be ((10⁷ / 10,000) + 200 + 100 + 100)256, which is 358,400.

So A will choose the path through B as the best path and will route traffic in that direction. Now, will A believe that the path through D is loop free? Let’s calculate the RDs to see.

Figure 1.5. A simple network, using real metrics

• The RD from B will be ((10⁷ / 10,000) + 100 + 100)256, which is 307,200.

• The RD from D will be ((10⁷ / 10,000) + 100 + 100)256, which is 307,200.

Is the RD through D less than the FD (the metric through the best path)? Yes, 307,200 is less than 332,800. So A will believe that the path through D is loop free and will mark it as a feasible successor (FS) in its topology table.

wow...