[c-nsp] do i *need* DFCs on the 6500?

[Geoffrey Pendery]

"congesting the crossbar fabric(sh fabric util) which is pretty
unlikely when you are talking a 24G linecard on a 40G fabric
connection"

Just a nitpick or two - the 6724 only has one channel of fabric, so
it's a 24G linecard with a 20G fabric connection, thus a slight
oversubscription.  If you're extremely rigorous about line-rate, leave
four of those ports unused.

But the DFC issue is more about Mpps than Mbps.  As Phil pointed out,
you could run them well under their line-rate, but with small packets,
and overwhelm the forwarding capacity of the PFC.


-Geoff


On Wed, Sep 2, 2009 at 5:59 PM, Ben Steele<illcritikz at gmail.com> wrote:
> Unless you are hitting a cam limit on any of your resources on your SUP(very
> possible if you are exporting netflow) OR you are congesting the crossbar
> fabric(sh fabric util) which is pretty unlikely when you are talking a 24G
> linecard on a 40G fabric connection then you probably won't see any
> difference putting a DFC on a 6724
>
> Remember these chassis are a hardware only based forwarding solution, so all
> your doing with a DFC is moving cam/asic resources off the sup, so in
> regards to your specific questions unless you have filled all your QoS
> queues on the sup you are going to see nothing more on the DFC, also the sup
> does (from memory) up to 100-200m pps in ipv6, I don't believe for a moment
> you are even remotely close to this, and the global ipv6 routing table is no
> where near the cam limit for that either, by the way is your SUP an XL? does
> the DFC's on the 10G's match the sup or have they fallen back to the lowest
> common configuration?
>
> "...or could it be that DFC's are only really useful to a particular
> deployment
> and I just *think* i need them?  ;-)" - I think you might be on the money
> here.
>
> If you give us the current utilization of your cam resources(from the sup)
> and the 6724 linecard throughput and what its functions
> are(netflow/qos/mac/acls etc) then we can tell you for sure.
>
> Ben
>
>
> On Wed, Sep 2, 2009 at 9:16 PM, Alan Buxey <A.L.M.Buxey at lboro.ac.uk> wrote:
>
>> hi,
>>
>> okay, from the background of I know what the DFC is and how it
>> operates etc... i know I want them - however, I need to justify
>> the upgrade/part cost to sort out a couple of 6500's.  in some of
>> our 6500's, the 10G blades have DFCs already...but several 6724's dont
>> (they just have CFC). ...as i said, I want them, but need to get
>> some management/funding buy-in - and they dont want the 'what it
>> does' information - they want some hard and fast facts that Cisco dont
>> sem to want to tell me ..... so, the question is
>>
>> 1) is there any way of showing the sup720 strain/utilisation...particularly
>> is there a way of showing DFC usage on the blades where we have them?
>>
>> 2) it offloads IPv6 and QoS - we're into both of those (and more so over
>> the
>> next year) - any particular insights into QoS performance/issues without
>> DFC ? any throughput figures for IPv6 ?
>>
>> (i know that with CFC we're limited to the backplane (32mpps?) and we get ~
>> 48mpps
>> per blade with DFC)
>>
>> ...or could it be that DFC's are only really useful to a particular
>> deployment
>> and I just *think* i need them?  ;-)
>>
>> alan
>> _______________________________________________
>> 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/
>>
> _______________________________________________
> 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/
>