Eigrp metric

[cosimodagostino]

Hello everyone, I start by saying all the formula for calculating the eigrp metric is clear except for one data. A simplified formula is 256 x [(10 ^ 7 / Least Bandwidth) + cumulative-delay]. I have not found anyone who explained it the only site I found was ciscopress but in a vague way it says: The EIGRP formula is based on the IGRP metric formula, except the output is multiplied by 256 to change the metric from 24 bits to 32 bits. But I don't understand why who can tell me?

I thank those who respond

[Kasun Bandara]

Hi,

you can find very good calculation explanation in below.

https://www.kwtrain.com/blog/understanding-eigrp-part-2

[cosimodagostino]

Hi I have already seen this site but it is not specified because it is multiplied by 256 at the beginning

[balaji.bandi]

There many example show how you can do, but if you looking deep it is Worth Looking RFC : (

5.6.  EIGRP Metric Calculations)

https://datatracker.ietf.org/doc/html/rfc7868

[Kasun Bandara]

Hi, that's the formula :D, you better check RFC then as balaji said.

[cosimodagostino]

So if you tell me that 256 is a conventional value decided without a reason, cisco could one day get up one morning and say from 256 we take it to 512. Is that what you want to tell me?

[Jon Marshall]

It is a Cisco proprietary routing protocol so in theory they could wake up one morning and change anything they wanted to if they were so inclined.

As for why I believe it was simply to give more granularity to the metrics in use.  

The IETF EIGRP draft written by the authors of EIGRP say basically that the metric was changed to "offer and enhance routing solutions for IGRP". 

Make of that what you will. 

Jon

[Joseph W. Doherty · Accepted solution]

". . . except the output is multiplied by 256 to change the metric from 24 bits to 32 bits."

As to why, I would presume to allow for greater precision and/or larger values (i.e. from 24 to 32 bits).  (I assume you realize multiplication by 256 shifts a binary number by 8 bits, i.e. the 24 to 32 bits.)  Also by moving to 32 bits, that generally aligns with various computer architectures (at that time) with "word" sizes of 32 bits.

You also noted this came about moving from IGRP to E(nhanced)IGRP.  The latter, I expect, allowed for, or expected, even larger networks than IGRP.

"So if you tell me that 256 is a conventional value decided without a reason, cisco could one day get up one morning and say from 256 we take it to 512. Is that what you want to tell me?"

Moving from 24 to 32 bits may make sense, for reasons noted above, however, perhaps at some point they (Cisco) will want to move from 32 bits to 64 bits.  Hmm, I haven't looked, but what's the metric formula in EIGRP IPv6?

[cosimodagostino]

But in fact my question was this to understand the reason for the initial 256 bits.Since many here have replied that it is part of the formula without giving a reason, I thought it was a convention.However, if it is not written anywhere, the motivation goes by intuition a bit like you did that maybe I think is the answer to my question

[paul driver]

Hello

---

@cosimodagostino wrote:

The EIGRP formula is based on the IGRP metric formula, except the output is multiplied by 256 to change the metric from 24 bits to 32 bits. But I don't understand why who can tell me?

---

My understanding for this specific value is based on those extra 8 bits added to the metric update and the reasoning is to add a more finer metric value for route path calculation.