[c-nsp] 802.3ad load distribution

[Eric Helm]

Speaking on this topic...
What would happen if there are 2 ports in the etherchannel that are each
hooked up via FastEthernet to a different wireless backhaul technology.?
The ports will be all linked at 100 Full, however the wireless links in
between the links are different and assymetrical. What happens to the
traffic being that the links are assymetrical, even though the switch
thinks they are 100 Full? What happens to the etherchannel traffic if
the wireless side of a single link drops but the ethernet side never drops?

Thanks,
Eric

Brady wrote:

> Hi,
> 
>>From an earlier post regarding etherchannel, which said:
> 
>  The answer is that you only get even distribution if you the # of
>  channels in the bundle is 2^n (n=1,2,3). Otherwise distribution might
>  look like
> 
>  For 3 port channel the load balancing would be 3:3:2
>  For 5 port channel the load balancing would be 2:2:2:1:1
>  For 6 port channel the load balancing would be 2:2:1:1:1:1
>  For 7 port channel the load balancing would be 2:1:1:1:1:1:1
> 
>  This is due to the hash algorithm which gives a value between 0 and 7.
> 
>  Arnold
> 
> ---
> I understand there are many ways to influence the loadbalancing (src-mac,
> dst-mac, src-ip, dst-ip, etc). But assuming the existing distribution is
> even at 2 links, is this information accurate for LACP and PAgP? Or would
> the results using LACP be significantly different? I'm wondering what type
> of loadbalancing I should expect to see when adding a 3rd link to the
> bundle. Any help is appreciated.
> 
> Regards,
> 
> Brady
> 
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>