[j-nsp] flow sampling: what packets are chosen?

[Michael Loftis]

On Thu, Jul 25, 2013 at 2:21 PM, Sebastian Wiesinger
<juniper-nsp at ml.karotte.org> wrote:
> * sthaug at nethelp.no <sthaug at nethelp.no> [2013-07-25 01:21]:
>> > "When using inline IPFIX the only valid rate is 1.  The option
>> > run-length isn't configurable, because there's no need to sample data
>> > from the perspective of the microcode in the Trio Lookup Block.  Every
>> > packet will be inspected and is subject to flow export."
>>
>> We're doing inline IPFIX with 1:100 sampling rate, and getting sensible
>> data. This is on an MPC-3D-16XGE-SFPP line card. Config snippet:
>
> I can confirm this, we use a higher sampling rate on MX80 and on the
> MPC-3D-16XGE-SFPP and on both we get valid data.


Good to hear, I remembered reading that mostly because it seemed so
wrong.  I don't think I'm missing any context from the book, and I
don't have any MPC/Trio based gear to test on myself.