At 10:33 PM 7/13/98 +0000, brad wrote:
>> BTW, if you do decide to use load balancing, I *highly* recommend
>> flow-switching. Otherwise your load balancing can get a bit... uneven. ;)
>> Even with flow-switching, you'll find it's not perfect. I hear you can do
>> per-packet load balancing with DCEF and not hit your main proc, but I
>> haven't tried it. Does anyone know if that will kill the VIP proc? And if
>> so, what rate will it handle?
>>
>I don't understand what the obsession with per-packet load balancing
>is and why people have such a hard time understanding how a 4 tuple
>load share hash(<source,sport,dest,destport>) could be uneven. In
>case you haven't noticed, you are selling a bursty service. Since you
>asserted that netflow switching is the best crisco has to offer, it
>is quite possible that you haven't noticed. As for dCEF, read the docs
>at cisco.. they tend to explain things so you don't have to ask questions.
A "4 tuple load share hash" can easily become uneven when you have a small
pipes - which is common if you are load balancing. Suppose I have two T1s
to a
POP out of which I have sold three T1s. I can easily get a significant load
(say 25 to 50%) on the two upstream T1s from two of the downstream T1s. Then
someone on the third T1 decides to download a big file and notices that he
never gets more than 50% of his bandwidth. Per-packet load balancing would
solve this problem.
As for the "busty product", in case you haven't noticed, you are not
running my
network or selling my products. My customers expect the bandwidth they paid
for all the time, not just when the 4 tuple balances properly. When they
connect to your network, you are welcome to give them bursty service if you
please.
Moving on to you claim that I asserted flow switching was the best "crisco"
had
to offer, I beg to differ. Even in the e-mail you quoted, I said quite
clearly
that "even with flow-switching, you'll find it's not perfect". Perhaps you
should learn to read yourself before posting?
Lastly, the docs are not always right, especially when it comes to performance
marks. I'm sorry if I thought I'd ask some people who might have tried it
real-world what their results are. You are, of course, welcome to believe
everything you read in cisco's docs and run your network in that way. I
choose
not to, but thanx for the suggestion.
>Frank "the font of all cisco knowledge" Rizzo
TTFN,
patrick
**************************************************************
Patrick W. Gilmore voice: +1-650-482-2840
Director of Operations, CCIE #2983 fax: +1-650-482-2844
PRIORI NETWORKS, INC.
<http://www.priori.net/>http://www.priori.net
"Tomorrow's Performance.... Today"
**************************************************************
This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:13 EDT