Eigrp Config....

Unanswered Question
Jul 13th, 2009

Hi....

I have two locations say "A" and "B"

And I have two 30MB lease line between these two location.I want to configure Eigrp between two location.I know about link failure using eigrp but can somebody tell how can i achive load balancing.

"A" --------- "B"

.........

Thanx in advance...

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 3 (1 ratings)
Loading.
Giuseppe Larosa Mon, 07/13/2009 - 03:31

Hello Arjun,

EIGRP will load balance over equal cost paths by default.

you can use

sh ip eigrp interface type x/y

sh ip eigrp topology

to check this

or simply

sh ip route x.x.x.x

you should see two paths

Hope to help

Giuseppe

arjunsawant Mon, 07/13/2009 - 05:31

Thanx for ur reply....

I have two 30 MB link with different service provider and is connected on same router at both locations.

Is there any difference for equal cost load balancing if i have two diff SP with same bandwidth.

further help is appreciated....

arjunsawant Tue, 07/14/2009 - 05:11

JOG-Tulip#sh ip route 10.1.0.3

Routing entry for 10.1.0.0/16

Known via "eigrp 111", distance 170, metric 28416, type external

Redistributing via eigrp 111

Last update from 10.5.0.149 on Vlan5, 04:51:14 ago

Routing Descriptor Blocks:

* 10.5.0.161, from 10.5.0.161, 04:51:14 ago, via Vlan6

Route metric is 28416, traffic share count is 1

Total delay is 110 microseconds, minimum bandwidth is 100000 Kbit

Reliability 255/255, minimum MTU 1500 bytes

Loading 13/255, Hops 1

10.5.0.149, from 10.5.0.149, 04:51:14 ago, via Vlan5

Route metric is 28416, traffic share count is 1

Total delay is 110 microseconds, minimum bandwidth is 100000 Kbit

Reliability 255/255, minimum MTU 1500 bytes

Loading 2/255, Hops 1

Can somebody tell me what is the meaning of *(Star) in * 10.5.0.161, from 10.5.0.161

Thanx...

Joseph W. Doherty Sat, 07/18/2009 - 15:45

BTW, came across info confirming usage of "*":

"For process-switching-load balancing is on a per-packet basis and the asterisk (*) points to the interface over which the next packet is sent.

For fast-switching-load balancing is on a per-destination basis and the asterisk (*) points to the interface over which the next destination-based flow is sent. "

from: http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a0080094820.shtml

Giuseppe Larosa Tue, 07/14/2009 - 05:46

Hello Arjiun,

>> Is there any difference for equal cost load balancing if i have two diff SP with same bandwidth.

EIGRP doesn't care who the provider is but it cares of declared bandwidth on L3 interfaces.

until the same set of EIGRP routes is received on both links load balancing will apply as it is in your show.

not sure the "*" might express a preference or rather the link '0' from the CEF load balancer point of view it is not part of EIGRP.

Hope to help

Giuseppe

arjunsawant Wed, 07/15/2009 - 20:17

Hi...

Pls find the attachment.

Is it write config?

Eigrp load balancing apprix 60-40%, but in my NMS it shows only 80-20%.

Pls guide me if I wrong.

Giuseppe Larosa Wed, 07/15/2009 - 22:11

Hello Arjun,

CEF load balancing decides what flows to send out link1 and out link2 regardless of traffic volume.

If you have few high volume traffic flows it is possible to see 80% / 20% on the links

Hope to help

Giuseppe

Joseph W. Doherty Thu, 07/16/2009 - 02:37

To confirm and expand upon what Giuseppe describes, exactly 50/50 routing load sharing is very rarely seen, especially short term. Even if you send just two flows, of exactly equal bandwidth demand, CEF might direct both to just one link.

What routing load balancing does, it "sort of" statistically balances flows, so the more you have, and over a longer and longer measured time period, usage normally approaches 50/50. Even then, and the reason I note "normally", CEF, if I recall correctly, deterministicly selects paths, so it's possible measured load will never be 50/50. (BTW, you can have a similar issue with load balancing on something like Etherchannel.)

The fact you see traffic across both links generally mean routing is, in fact, using both paths.

One method to improve actual load balancing is using OER/PfR technology. (NB: Until [very] recently, EIGRP wasn't supported [until PIRO feature].)

Another approach to improve actual load balance is to use some form of link bonding technology that alternates packets across multiple links (somewhat like CEF packet-by-packet) but preserves packet sequencing, e.g. MLPPP.

Again, though, if you're seeing 80/20 in your NMS, especially short term, it's very likely routing load balancing is working as it's supposed to.

Joseph W. Doherty Mon, 07/13/2009 - 03:35

Not 100% sure with EIGRP w/o reviewing its docs, but I believe it will load balance across 2 equal cost paths by default.

If the paths between your sites use different routers, or if metrics make the paths look different, or if there are more than 4(?) paths, then load balancing becomes more involved.

Actions

This Discussion