Unanswered Question
Sep 14th, 2007


I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
Kevin Dorrell Fri, 09/14/2007 - 04:42

The "administrative distance" and the "EIGRP metric" are two completely different things.

The AD (Administrative Distance) is a way of arbitrating between routes when you have more than one routing protocol in the router, e.g. EIGRP and BGP, or OSPF and static. The AD tells you the pecking order of the routes learned from the variuous sources. For example, static routes (AD=1) have priority over EIGRP internal routes (AD=90), which have priority over OSPF routes (AD=110). etc.

The EIGRP metric arbitrates between routes learned by the EIGRP routing process from different sources. The EIGRP metric consists of 5 variables (bandwidth, delay, reliability, load, and MTU) of which usually only bandwidth and matric are used. The variables are massaged together in each individual router to form a composite metric, but they are passed from router to router as individual variables.

To see these metric values, do show ip route on an EIGRP route.

Kevin Dorrell


P.S. Please don't shout - please use lower case where appropriate. It would make your posting a lot more readable.

pratap_deysarkar Fri, 09/14/2007 - 04:56

Sir, thank you very much for your reply. But I wanted to know about advertised distance not about administrative distance.Advertised distance (AD) which is used to select feasible successor(FS).

Kevin Dorrell Fri, 09/14/2007 - 05:49

Sorry. I should have read the question better.

The Feasible Distance (FD) is the composite metric of the best EIGRP route that I have for a certain destination.

In EIGRP, the router can also keep a record of a "next best" route, known as a "successor route". It is here that the Reported Distance (RD), also known as the Advertised Distance (AD) comes into play. We want to make sure that this "next best" routes are not actually loops, or are not dependent on the primary route. This is the loop prevention mechanism of EIGRP.

I choose the route that has the lowest metric seen from where I am. Each router also tells me the metric from where he is - this is the RD or AD. The rule is that I can believe any route from a router that is closer to the destination than I am. That is, I discard any routes that have an RD greater than my FD.

So, the simple answer is that the RD (or AD) of a route is the metric of that route at the next-hop router.

Kevin Dorrell


paul.matthews Fri, 09/14/2007 - 05:46

Please be careful - I note Kevin's response and your reply AD is normally used to refer to administartive distance.

Yes there is a formula - this is a opportunity to mention Jeff Doyle - routing TCP/IP Vol 1. I have not seen edition two which is apparently a revision.

The advertised distance is the router's own metric - as with most DV protocols, even if that is propagated in component form rather than the final value. EIGRP can be tuned in how the metric is calculated.

By default, the two parameters used are total delay and minimum bandwidth.

IGRP has the metric (default) as

(10,000,000/the minimum bandwidth)+(sum of delays/10)

EIGRP multiplies bandwidth and the sum of delays by 256. (P329)

e.g min band =10M, delay(sum)=250 IGRP metric=1020, EIGRP =261120.

So if router A has a network where the smallest link on the path to it is 10M, and the sum of delays is 250, the router's own metric in eigrp would br 261120. As router A advertised that to router B the advertised distance would be 261120.

This is quite important in loop avoidance and quick convergence in EIGRP.

The basic premise with a fasible sucessor is that the metric is higher than the current best I am getting, but is lower than mine, therefore I can use it if the primary route is lost, because the metric is lower than mine, so does not return to me at any point.

pratap_deysarkar Fri, 09/14/2007 - 06:33

Thank you Sir,it is now more clear to me.

I want to make it clear a bit more,

say 4 routers(A,B,C,D) are connected back to


The metric of the path A to D is the FD..

what is the AD(Advertised Dis)in his case,

the metric between B and D or C and D ?

paul.matthews Fri, 09/14/2007 - 06:47

assuming they are connected in a line:


and the target network is attached To A, the AD from A to B will be the metric A has for the connected network.

B will add the delay it has for the link A-B, and look at the bandwidth between A-B and calculate the metric and install the route in the routing table.

B will advertise that to C and between B&C the Advertised Distance will be the nummber in the routing table of B.

C will see the advert, and look at the bandwidth B-C, and add the delay to the total del, calculate the route and add to the routing table. It will advertise to D, and the AD between C&D will be the number in C's routing table.

The Advertised Distance will get higher the further away from where the network is originated, even if all links are the same bandwidth, delay gets incremented at each step.

pratap_deysarkar Fri, 09/14/2007 - 07:24

Thanx a lot Sir, i m asking one thing more.

if metric between

A & B = 2

B & C = 3

C & D = 4

what will be metric between A & C ?

Will it be 2 + 3 = 5 ?

Metric of a route and Feasible Distance of a route are same?

paul.matthews Fri, 09/14/2007 - 07:40

the metric on EIGRP is a compound one, so doen not necessarily increment tidily like that.

If the delay a-b = 2 and the delay b-c is 3, then the total delay will be 5, but if the bandwidth B-C is greater than a-b, then the bandwidth bit will be unchanged, only the delay component is cumulative, bandwidth is the lowest on the path. If all links in a neetwork are the same bandwitdh, the cumulative delay will effectively be the metric.

On any router, the FD is the lowest metric seen to a destination, so on router B the actual metric on the route installed in the routing table is the FD.

Instead of a line, consider a triangle. I will ignore the real metrics in this case. I will drop to delay only cumulative with no maths!

We have a triangle - 1, 2 & 3 and we are thinking about a network via R1, but from the perspective of router 3.

On router 1, The route has a metric of, say, 200. Router 1 will advertise that route to R2 & R3 with an advertised distance of 200.

R2 & R3 will each add their delay - let's say R2 has a delay of 25 on the link, and R3 has a delay of 50

They will then have R2 - 225, R3 - 250

R2 now has an FD of 225 for the route, and R3 250.

Being DV R2 & R3 will advertise to each other - R2 will advertise at 225, R3 will advertise at 250.

Assume the link R2 has a delay of 50. If accepted, the route via R3 From R2 would have a metric of 300, and via R2 from R3 will be 275.

Both of these are higher than the current metric, so will not be considered for the routing table. We now look at the AD. From R2s point of view, FD = 225, AD from R3 = 250 - not a feasible successor, but on R3 FD = 250, AD = 225 so the route via R2 is a fasible successor.

pratap_deysarkar Sat, 09/15/2007 - 23:23

Thanx a lot Sir, onthat day i was not able to reply you.it is now clear to me.I got confused coz, this line "If all links in a neetwork are the same bandwitdh, the cumulative delay will effectively be the metric." is not mentioned in cisco press BSCI book.

thank u once again.


This Discussion