AFAIK, MLPPP does do fragmentation, distribution and reassembly over
the links, effectivelly doing IM over any supported link
(IMA stands for Inverse multiplexing atm).
That's why it used to be the case that if you had voip, cisco recomended
to configure MLPPP even on only one link.
And yes, you should be able to test out this with just a performance
check between two machines. Without MLPP, cef will use only one link,
with MLPPP you should get the compounded BW.
-Carlos
Christopher Copley @ 8/01/2010 17:40 -0300 dixit:
> Guys,
> Sorry for the OT, but wanted to push this out and get some thoughts on it.
> We just got a new account mgr from our telco, and he is tiring to get us to
> go to IMA over our current MLPPP for local termination to our MPLS. I
> have argued with with that if I have 3 IMA T1's and 3 MLPPP T1's that the
> MLPPP will get better performance b/c of less over head than with the ATM
> IMA links.
> He is tiring to tell me that MLPPP is not really bonded that it basically is
> doing per-packet load balancing per T1 and you will never get the full use
> of the T1's.
> I personally feel he is tiring to scam our upper mgt to getting more
> expensive CE routers with the ATM AIM cards. I personally can not find any
> doc's on this subject out there, and don't have a good way to test this.
> Any one have any thoughts?
>
> Thanks
> Chris
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
>
>
>
>
>
-- Carlos G Mendioroz <tron_at_huapi.ba.ar> LW7 EQI Argentina Blogs and organic groups at http://www.ccie.netReceived on Fri Jan 08 2010 - 19:47:39 ART
This archive was generated by hypermail 2.2.0 : Thu Feb 04 2010 - 20:28:41 ART