|
|
|
| What is the metric used for EIGRP?
The Cisco press study guide I am using says the default is:
[(10,000,000/smallest bandwidth kbps)+delay]*256
and then goes on to list five metrics
K1 Bandwidth
K2 Loading
K3 Delay
K4 Reliabilty
K5 MTU
I am just a bit confused. Does this mean that any combination of the 'K' values can be manually configured, but the default is as per the formula given above.
Any help/clarification would be appreciated.
Cheers
Hippo
 | |
| Yankee 2002-09-24, 3:06 am |
| Yeppers, that is what it means!
Yankee | |
| dmaftei 2002-09-24, 6:09 am |
| The default formula is obtained from the general formula:
metric = [K1 * bandwidth + (K2 * bandwidth) / (256 - load) + K3 * delay] * [K5 / (reliability + K4)]
by giving default values to Ks.
(No, I don't know this formula off the top of my head; I looked it up at http://www.cisco.com/warp/public/103/eigrp1.html ) | |
|
| Thanks guys
dmaftei, I will take that formula away and find a sample router output and see how it pans out. However, it does seem overly complicated when compared to the OSPF metric.
Hippo
| |
| necrophantasm 2002-09-24, 7:14 pm |
| Just remember the default metrics are bandwidth and delay, although EIGRP can use the others if configured to do so. I can't speak for CCNP, but all of the tests I have taken so far have only wanted me to know which were defaults, not the actual formulas involved in determining preferred routes.. You guys tell me if I should be memorizing that one (or if I will actually ever have to know that in a work situation). | |
| Yankee 2002-09-25, 3:10 am |
| Me thinks Hippo wants to understand EIGRP and is not as concerned with what is on the test. In the real world you need to know which metrics "your" EIGRP is using so you may understand why it selected a particular route and not another.
Yankee | |
|
| Yankee
You are so right, and I hope I don't come across a real-world EIGRP network.
Hippo
| |
| Yankee 2002-09-25, 6:28 pm |
| Don't sweat it Hippo.... there is no real need that I am aware of to change from the default metrics and if you continue reading you will learn only bw and delay changes cause an immediate change in the metric. Load and the rest of the crud wait so they are unnecessary in real networks as far as I be concerned.
Yankee | |
| Widgewaam 2002-09-25, 8:04 pm |
| Is it not true that the only major difference between IGRP and EIGRP is that the EIGRP metric is mulitplied by 256? This allows more granularity, but is there any other difference between the two? | |
| Yankee 2002-09-26, 3:08 am |
| no there are great differences between the two protocols, however the only difference between their metrics is what you stated.
Yankee | |
| edmonds_robert 2002-09-26, 9:51 pm |
| quote:
Originally posted by Widgewaam
Is it not true that the only major difference between IGRP and EIGRP is that the EIGRP metric is mulitplied by 256? This allows more granularity, but is there any other difference between the two?
IGRP is a classful routing protocol and therefore does not send subnet mask information with its routing updates. This means that it also doesn't support VLSM (Variable Length Subnet Masks). This severely limits the scalability of IGRP in an enterprise networking environment.
EIGRP is a classless routing protocol and does support VLSM. Although I wouldn't consider it nearly as scalable as OSPF, it is much better than IGRP or RIP (version 1 or 2).
I run EIGRP exclusively on my network, and I really like it. It converges faster than you can get back to a command line to ping to check the status. Most other routing protocols take at least a few more seconds, if not minutes (in the case or RIP) to converge. | |
| Widgewaam 2002-09-27, 12:00 pm |
| Oh yeah, I forgot that bit about VLSM. No pun intended. |
|
|
|