Re: PAgP Learn Method and Port-channel Loadbalancing

From: Miroslav Kosut <miroslav.kosut_at_me.com>
Date: Tue, 10 Nov 2009 21:48:55 +0100

Thank you, actually that is ok.

However what if we have configured pagp learn-method to be a "physical-
port" learning. Then a destination mac address is already in the table
- on physical interface, so no load balancing is necessary (port is
already chosen, because frame was received on that etherchannel, and
learned on the physical-port). The load balancing wouldn't be ignored
if learn-method was "aggregate-port" learning - then actual interface
would have been calculated using a hashing algorithm. This is how I
understand it.

Yes, hashing is pretty easy:

1. Loadbalancing based on one field (not a pair) such as src-mac / dst-
mac / src-ip / dst-ip / .. is done using last N bits of that address.
(num_of_ports = 2^N)
2. Loadbalancing based on a pair of addresses such as s src-dst-mac /
src-dst-ip / ... is done using XOR operation on last N bits of both
addresses..

Your feedback is appreciated!

Regards,
Miroslav Kosut

On Nov 10, 2009, at 9:37 PM, S Malik wrote:

> Dstn-mac- add based LB means that switch will calculate the hash
> based on destination mac-add and select he actual port accordning to
> the hash result and # of ports invloved int the ether-channel.
> Once a ether-channel is formed then then selection of actual port
> depends upon the Hash result and # of ports in the channel.
>
> Any correction is welcomed.
>
> On Tue, Nov 10, 2009 at 3:09 PM, Miroslav Kosut
> <miroslav.kosut_at_me.com> wrote:
> Dear group,
>
> a quick question to confirm my understanding. Suppose I configure
> the following on the same etherchannel:
>
> 1. PAgP learning method to be a "physical-port learning"
> 2. Port-channel loadbalancing to be based on a "destination mac
> address"
>
> Then, am I right when I am saying that load-balancing method is
> ignored when a destination mac address is already learned (on a
> physical-port) ?
>
> My understanding:
> Since the mac is already learned on a physical-port, there is no
> need to calculate the outgoing member interface (load-balancing).
> Of course, if the mac address hasn't been learned, frames would be
> unicast-flooded. Then, load balancing IS performed and the outgoing
> interface is chosen.
>
> Regards,
> Miroslav Kosut
>
>
> 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 Tue Nov 10 2009 - 21:48:55 ART

This archive was generated by hypermail 2.2.0 : Tue Dec 01 2009 - 06:36:28 ART