cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
4713
Views
0
Helpful
6
Replies

EIGRP and composite metric calculation

g.peart
Level 1
Level 1

I have a summary address of 172.16.0.0/16  configure on a router interface

which is being advertised to another network segment.

I installed a TenGig module in my router to link another segment

and I noticed that the composite metric value changed from 2816/0 to 512/0

The summary address is configured on a 1G interface not the TenGig so why would

the metric change, since the minimum bandwidth on the router is 1G and the EIGRP formula

doesn't seem to obey the rules because its taking the maximum bandwidth on the router.

The same thing happened when I created a 2G layer 3 etherchannel and the

composite metric changed to 1536/0.

Do differnet rules apply when the Null interface is involved

thanks.

Router# sh ip eigrp top 172.16.0.0 255.255.0.0

IP-EIGRP (AS-1): Topology entry for 172.16.0.0/16

State is Passive, Query origin flag is 1, 1 Successor(s), FD is 512

Routing Descriptor Blocks:

0.0.0.0 (Null0), from 0.0.0.0, send flag is 0x0

Composite metric is (512/0), Route is Internal

Vector metric:

Minimum bandwidth is 10000000 Kbit

Total delay is 10 microseconds

Reliability is 255/255

Load is 0/255

Minimum MTU is 1500

Hop count is 0

1 Accepted Solution

Accepted Solutions

Consider the following topology:

R1(10gig)-10.1---(10.10.10.0/24)----10.2-(10gig)R2(faEth)-1.1-----(192.168.1.0/24)---1.2-(faEth)R3

                                                                      |

                                                                      |

                                                                      |--(gig)-.11.1------(10.10.11.0/24)-----11.2-(gig)R4

R1, R2, R3 and R4 are EIGRP neighbors.

At R2's fastethernet interface, i configure a summary route for 10.0.0.0/8 , for R3.

First, it must be noted that a summary route will not sent from R2 to R3 if R2 doesnot have a more specific subnet in it's routing table within the summary subnet it advertizes - which is 10.0.0.0/8 in this case. Since R2 has 2 interfaces with subnets - 10.10.10.0/24 and 10.10.11.0/24 which are a subset of 10.0.0.0/8, R2 will advertize the summary to R3.

Now, the question arises - with what metric will the summary route - 10.0.0.0/8 be advertised to R3 ?

The answer is what Alain pointed out. Cisco engineers have chosen that the best metric for routes with subnets which are a subset of the summary subnet will be the metric of the summary route.

In this case R2 has both 10.10.10.0/24 and 10.10.11.0/24  are a subset of the 10.0.0.0/8 summary subnet and they show up as connected routes in R2's routing table.

Obviously, 10.10.10.0/24 will have a better metric than 10.10.11.0/24 because 10.10.10.0/24 is on a TenGig interface, whilst the other is on a Gig interface. For ex consider metric for 10.10.10.0/24 = 100, and metric for 10.10.11.0/24 = 500

The metric for the summary route 10.0.0.0/8 will be the best metric - between 10.10.10.0/24 and 10.10.11.0/24. - between 100 and 500

Since 10.10.10.0/24 has a better metric, the metric of the summary route 10.0.0.0/8 will be 100.

What happens, if in future, I connect another router R5 via a 100 Gig link to R2, and with a subnet 10.10.12.0/24 ? Since the metric of 100Gig will be better than 10Gig - let us say it is 10, the metric of the summary route will change to 10. (best among smaller subnets). Of course the subnet on the 100 gig will have to be within the summary subnet. 10.10.12.0/24 is within 10.0.0.0/8. If I put another subnet , but not part of 10.0.0.0/8, it will not affect the metric (obviously )

Is this behaviour limited to connected subnets alone ?. No

This behavior is for a route learnt in any manner through EIGRP - connected, redistributed from static or other protocols etc, but it needs to be present in the routing table

HTH

View solution in original post

6 Replies 6

cadet alain
VIP Alumni
VIP Alumni

Hi,

the EIGRP metric of a summary route equals the best metric of the routes belonging to that summary.

Regards.

Alain.

Don't forget to rate helpful posts.

Just to be sure that I understand, the summary is being calculated using the

bandwidth and delay values of the best metric on the router.

The 10G layer 3 interface on the router has a starting composite metric of 512/0

for the particular route it is configured for, and the summary address metric is based upon

those values.

So if I moved the 10G to a different router that router will be the starting point for the calculation

of summary address metric and the previous router will calculate the new value from an update

based upon its minimum bandwidth path to that route + the delay value on the incoming interface.

Example:

The two routers are connected by 1G layer-3 link, with Router A having 10G layer 3 interface configured.

Router A 172.16.0.0/16 (512/0)

Router B 172.16.0.0/16  [10^7 /10^6] * 256 + ( [1+1] *256) (3072/512)

Consider the following topology:

R1(10gig)-10.1---(10.10.10.0/24)----10.2-(10gig)R2(faEth)-1.1-----(192.168.1.0/24)---1.2-(faEth)R3

                                                                      |

                                                                      |

                                                                      |--(gig)-.11.1------(10.10.11.0/24)-----11.2-(gig)R4

R1, R2, R3 and R4 are EIGRP neighbors.

At R2's fastethernet interface, i configure a summary route for 10.0.0.0/8 , for R3.

First, it must be noted that a summary route will not sent from R2 to R3 if R2 doesnot have a more specific subnet in it's routing table within the summary subnet it advertizes - which is 10.0.0.0/8 in this case. Since R2 has 2 interfaces with subnets - 10.10.10.0/24 and 10.10.11.0/24 which are a subset of 10.0.0.0/8, R2 will advertize the summary to R3.

Now, the question arises - with what metric will the summary route - 10.0.0.0/8 be advertised to R3 ?

The answer is what Alain pointed out. Cisco engineers have chosen that the best metric for routes with subnets which are a subset of the summary subnet will be the metric of the summary route.

In this case R2 has both 10.10.10.0/24 and 10.10.11.0/24  are a subset of the 10.0.0.0/8 summary subnet and they show up as connected routes in R2's routing table.

Obviously, 10.10.10.0/24 will have a better metric than 10.10.11.0/24 because 10.10.10.0/24 is on a TenGig interface, whilst the other is on a Gig interface. For ex consider metric for 10.10.10.0/24 = 100, and metric for 10.10.11.0/24 = 500

The metric for the summary route 10.0.0.0/8 will be the best metric - between 10.10.10.0/24 and 10.10.11.0/24. - between 100 and 500

Since 10.10.10.0/24 has a better metric, the metric of the summary route 10.0.0.0/8 will be 100.

What happens, if in future, I connect another router R5 via a 100 Gig link to R2, and with a subnet 10.10.12.0/24 ? Since the metric of 100Gig will be better than 10Gig - let us say it is 10, the metric of the summary route will change to 10. (best among smaller subnets). Of course the subnet on the 100 gig will have to be within the summary subnet. 10.10.12.0/24 is within 10.0.0.0/8. If I put another subnet , but not part of 10.0.0.0/8, it will not affect the metric (obviously )

Is this behaviour limited to connected subnets alone ?. No

This behavior is for a route learnt in any manner through EIGRP - connected, redistributed from static or other protocols etc, but it needs to be present in the routing table

HTH

Perfect, I understand what's going on now, did a bench test changed the routes around

and introduced a new major network 172.17.0.0/16 with a summary address on an interface

and it worked out has outlined.

thanks

Hi,

You can summarize only if you've got the summary routes in your RIB so lets suppose you know of 172.16.2.0/24, 172.16.3.0/24 and 172.16.4.0/24

and you want to summarize with a 172.16.0.0/16.

In your rib you have 172.16.2.0 (90/250012)- 172.16.3.0 ( 90/200012) - 172.16.4.0 ( 90/180020)

then your summary will be advertised as  172.16.0.0 ( 90/ 180020) and if that 172.16.4.0 route dies then while you still got one route the summary is still advertised but with 200012 as feasible distance now.

Regards.

Alain.

Don't forget to rate helpful posts.

Thanks for the extra input cadetalain, helped it all fall into place.