[c-nsp] Cisco and microbursts

[Saku Ytti]

On 19 May 2016 at 14:14, Pavel Skovajsa <pavel.skovajsa at gmail.com> wrote:

Hey,

> Maybe they are just trying to uncover the fact that for example old 2621
> router has slow clock on its ASIC (in this case mips cpu) but it actually is
> capable of having multiple of  FastEthernet interfaces. They do not work
> really great, but possible :)
>
> If you speed up that 'asic' things are suddenly great! Who would have
> though? So easy...

Yes, this is the 'lookup starved/congested' issue, not microburst. So
even if this was the case, it wouldn't be interesting data when trying
to solve microburst issue.

This particular box is rated for 65Mpps lookup performance, the test
was using <0.8Mpps.

-- 
  ++ytti