Re: [squid-users] squid deployment in 6Gbit network with tproxy as L2 bridge.

On 08/23/2013 03:46 PM, Pawel Mojski wrote:
> I hope you understood any of it :)
>
> So, at the end of the day balancert have no balancing software
> installed, everything is done on the kernel-space. And each scanner
> received only part of the traffic so no problems with ip stock limits.
>
> Regards;
> Pawel Mojski
Hey there,

I do think I understood couple things but I see it from another angle.

Kernel level is not always the smartest way of doing things..
Many developers are just familiar with the lower levels languages and
prefer them.
When you have 5 ghz for example and you do have 4 stages of a CPU cycle
with let say 32\64 cores you will most-likely wont have any trouble
moving 1GB on each machine...
Once we understand that we have so much power in one machine we just
leave the lower levels programmers to do their best and move to the
upper level to make sure that the client will get the job done without
any reliability on the kernel level.
There is so much on the kernel level that is Going ON that an
application can just run and the basic example for the stability of the
kernel level is the CentOS and Debian kernel examples that just works
out of the box like that for *ages*.
We are not talking about some embedded system and more about a first
class hardware system.
Squid aims to run on both but you just cannot move 4 Tons of steel on a
simple tiny bike..

I must say that this kind of testings are a really good thing And I will
try to look over it.

Regards,
Eliezer
Received on Mon Aug 26 2013 - 22:07:52 MDT