[c-nsp] Gigabit EtherChannel

[Oliver Boehmer (oboehmer)]

Joe Maimon <mailto:jmaimon at ttec.com> wrote on Thursday, June 29, 2006
2:43 PM:

> Oliver Boehmer (oboehmer) wrote:
> 
>> Jon Lewis <mailto:jlewis at lewis.org> wrote on Wednesday, June 28, 2006
>>> If the bandwidth is (or becomes) much higher than what a single GEIP
>>> can handle, wouldn't it be a bad idea unless the goal is just
>>> redundancy? Won't the single GEIP+ on its own handle about as much
>>> traffic as a GEIP & GEIP+ etherchannel can handle?
>> 
>> 
>> I did not talk about performance of the whole thing. a GEIP or GEIP+
>> on a 7500 will not do line-rate (given IMIX) anyway, so I assumed
>> that this is targeted for redundancy reason, rather than load..  I
>> guess you are right, a GEIP+ by itself will likely provide similar
>> performance compared to a GEIP&GEIP+ channel at the end..
>> 
> 
> Wouldnt spreading the load across multiple vip4-80 effectively
> increase the available throughput? SO in that sense, yes this would be
a
> performance benefit.

Yes, of course, two GEIP+ bundled in a channel would increase the
overall performance. But mixed with a GEIP, it might or might not..

	oli