[c-nsp] T1 Bonding with PA-MC-T3

[Andrew W. Smith]

Nick Voth wrote:
>> [Deleted]
>>
>>     
>>> SO, is there any way to accomplish T1 bonding with that existing DS3
>>> card or am I just stuck?
>>>       
>> Nick,
>> Is there any specific reason why you need MLPPP? How about just
>> using per-packet load balancing? We do that on our PA-MC-T3 via the
>> following:
>> [snipped]
>
> Greg,
>
> To be honest, I've never heard of that. I looked up the command reference
> and it seems understandable, but I'm not totally sure what it's doing yet.
> There's per-packet and per-destination load-sharing and it's not really
> clear what would be appropriate for my situation.
>
> What I need to accomplish is to effectively deliver a 3Mb pipe to an end
> user. It's my understanding that with MLPPP we can accomplish that. In other
> words, the individual T1's are actually bonded so transferred data will come
> through at 3Mb instead of 1.5. Can the load-sharing technique accomplish the
> same thing?
>
> Also, this link will have VoIP on it. I see you have a voice-policy. Have
> you had any VoIP issues using the load-sharing technique?
>
> Thanks very much!
>
> -Nick
>
>
> _______________________________________________
> 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/
>
>
>   
In my experience, avoiding MLPPP on serials was usually used when there 
were significant differences between the links in delay  or reliability, 
or when one of the devices couldn't handle the extra CPU load that MLPPP 
brought compared to easier load balancing techniques.

MLPPP is also generally much better than per-destination load balancing 
when any individual destination is utilizing a significant amount of 
traffic compared to the speed of any individual link in the bundle.

Andrew W. Smith
Sr. Network Engineer
Internet America, Inc.
awsmith at airmail.net
(w)713-968-2506
(c)832-457-3138