Re: EIGRP unequal metric LB

Hi,

Thanks for the response.
I am trying to figure out the the maths required to achieve the last
painful option :(

Here is the output of what I am working with

Rack1R4(config-if)#do sh ip route 222.22.2.0
Routing entry for 222.22.2.0/24
  Known via "eigrp 1024", distance 170, metric 156928, type external
  Redistributing via eigrp 1024
  Last update from 174.1.145.1 on Serial0/0/0, 00:00:25 ago
  Routing Descriptor Blocks:
    174.1.145.5, from 174.1.145.5, 00:00:25 ago, via Serial0/0/0
      Route metric is 567552, traffic share count is 11
      Total delay is 21170 microseconds, minimum bandwidth is 100000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1
    174.1.145.1, from 174.1.145.1, 00:00:25 ago, via Serial0/0/0
      Route metric is 567808, traffic share count is 11
      Total delay is 21180 microseconds, minimum bandwidth is 100000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1
  * 174.1.45.5, from 174.1.45.5, 00:00:25 ago, via FastEthernet0/1
      Route metric is 156928, traffic share count is 40
      Total delay is 5130 microseconds, minimum bandwidth is 100000 Kbit
      Reliability 255/255, minimum MTU 1500 bytes
      Loading 1/255, Hops 1

Rack1R4(config-if)#do sh ip ei topo 222.22.2.0
IP-EIGRP (AS 1024): Topology entry for 222.22.2.0/24
  State is Passive, Query origin flag is 1, 1 Successor(s), FD is 152064
  Routing Descriptor Blocks:
  174.1.45.5 (FastEthernet0/1), from 174.1.45.5, Send flag is 0x0
      Composite metric is (156928/25600), Route is External
      Vector metric:
        Minimum bandwidth is 100000 Kbit
        Total delay is 5130 microseconds
        Reliability is 255/255
        Load is 1/255
        Minimum MTU is 1500
        Hop count is 1
      External data:
        Originating router is 150.1.5.5
        AS number of route is 0
        External protocol is RIP, external metric is 7
        Administrator tag is 0 (0x00000000)
  174.1.145.5 (Serial0/0/0), from 174.1.145.5, Send flag is 0x0
      Composite metric is (567552/25600), Route is External
      Vector metric:
        Minimum bandwidth is 100000 Kbit
        Total delay is 21170 microseconds
        Reliability is 255/255
        Load is 1/255
        Minimum MTU is 1500
        Hop count is 1
      External data:
        Originating router is 150.1.5.5
        AS number of route is 0
        External protocol is RIP, external metric is 7
        Administrator tag is 0 (0x00000000)
  174.1.145.1 (Serial0/0/0), from 174.1.145.1, Send flag is 0x0
      Composite metric is (567808/25856), Route is External
      Vector metric:
        Minimum bandwidth is 100000 Kbit
        Total delay is 21180 microseconds
        Reliability is 255/255
        Load is 1/255
        Minimum MTU is 1500
        Hop count is 1
      External data:
        Originating router is 150.1.1.1
        AS number of route is 1
        External protocol is OSPF, external metric is 20
        Administrator tag is 0 (0x00000000)

From:
Andy Reid <ccie_at_reid.it>
To:
michael.botha_at_gm.com
Cc:
ccielab_at_groupstudy.com
Date:
2009/07/30 10:27 AM
Subject:
Re: EIGRP unequal metric LB

Hi Michael,

I am going to need more information.
First, what is the network and router from which you are attempting to
perform unequal cost load balancing.
Then, on that router, what is the output of "show ip eigrp AS topology
IP/MASK"
Is variance ruled out, if not, then I would be using the variance X
command (as long as the advertised distance of the worst route is lower
than the feasible distance).
If it isn't an even multiplier (which it is here - x 4) then you can
change the maximum paths also.

Note that if you are not allowed to use the variance command then you
can use an offset list to manipulate and add to the metric value (from
the inbound interface) - though you cant use it to subtract from the
metric value.

The last option would be to manipulate bandwidth and delay values - this
is more painful as it is slow and hard to get the right balance.

regards Andy

michael.botha_at_gm.com wrote:
> Hi,
>
> I am really battling to calculate the delay required to get unequal cost
LB
> ratio's correct.
>
> I have two routes to a destination prefix.
>
> One via a frame-relay link and the other via an ethernet segment.
>
> I need to balance the traffic with a ration of 4:1.
>
> By setting the delay I manage to get a ration of 40:11 which is pretty
close to
> 4:1 but not exactly. I then up the delay on one interface and turn it
down on
> the other and I get a ratio of 38:10.
>
> Anyways - bottom line is. I am missing the maths here to get these ratio
100%
> spot on.
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html

Blogs and organic groups at http://www.ccie.net
Received on Thu Jul 30 2009 - 10:52:15 ART