On Tue, 1 Nov 2011 08:55:07 -0400
Nico Kadel-Garcia wrote:
> On Tue, Nov 1, 2011 at 4:10 AM, Gregory Edigarov
> wrote:
> > On Tue, 1 Nov 2011 08:53:46 +0100
> > "Bret S. Lambert" wrote:
> >
> >> On Tue, Nov 01, 2011 at 09:47:35AM +0200, Gregory Edigarov wrote:
> >> > On Tue, 1 Nov 2011 11:17:56
ZZ Wave writes:
> What solution should be used for traffic shaping on real-life, "production"
> gateways with tens and hundreds users? PF queues seem to be too
> "userspace"-ish and CPU consuming.
PF setups with various altq disciplines are serving sites with larger
user bases than that.
If i
ZZ Wave writes:
> For example, in FreeBSD there is "slow" pf in userspace and "fast"
> kernel-level netgraph.
Wow, I can scarcely imagine a single sentence that reveals more
thoroughly and conclusively how little familiarity you have with any of
the systems you mention.
Hint: both pf and netgra
On Tue, Nov 1, 2011 at 4:10 AM, Gregory Edigarov
wrote:
> On Tue, 1 Nov 2011 08:53:46 +0100
> "Bret S. Lambert" wrote:
>
>> On Tue, Nov 01, 2011 at 09:47:35AM +0200, Gregory Edigarov wrote:
>> > On Tue, 1 Nov 2011 11:17:56 +0400
>> > ZZ Wave wrote:
>> >
>> > > What solution should be used for tr
On Tue, Nov 1, 2011 at 9:26 AM, ZZ Wave wrote:
> For example, in FreeBSD there is "slow" pf in userspace and "fast"
> kernel-level netgraph.
And what has this to do with OpenBSD?
ah, you mean nat? In OpenBSD all firewall functions (uhmm, almost all,
to be technically correct, in the presence of [t]ftp-proxy) i.e. packet
filtering, NAT, shaping are done on the kernel level.
On Tue, 1 Nov 2011 12:26:30 +0400
ZZ Wave wrote:
> For example, in FreeBSD there is "slow" pf in
On Tue, Nov 01, 2011 at 12:26:30PM +0400, ZZ Wave wrote:
> For example, in FreeBSD there is "slow" pf in userspace and "fast"
> kernel-level netgraph.
*headasplode*
>
> 2011/11/1 Gregory Edigarov
>
> > On Tue, 1 Nov 2011 11:17:56 +0400
> > ZZ Wave wrote:
> >
> > > What solution should be used
On Tue, Nov 01, 2011 at 12:26:30PM +0400, ZZ Wave wrote:
| For example, in FreeBSD there is "slow" pf in userspace and "fast"
| kernel-level netgraph.
This isn't a FreeBSD list. This is OpenBSD - pf is in the kernel.
And besides .. do you think the cpu runs slower when it's executing
userland cod
For example, in FreeBSD there is "slow" pf in userspace and "fast"
kernel-level netgraph.
2011/11/1 Gregory Edigarov
> On Tue, 1 Nov 2011 11:17:56 +0400
> ZZ Wave wrote:
>
> > What solution should be used for traffic shaping on real-life,
> > "production" gateways with tens and hundreds users?
On Tue, 1 Nov 2011 08:53:46 +0100
"Bret S. Lambert" wrote:
> On Tue, Nov 01, 2011 at 09:47:35AM +0200, Gregory Edigarov wrote:
> > On Tue, 1 Nov 2011 11:17:56 +0400
> > ZZ Wave wrote:
> >
> > > What solution should be used for traffic shaping on real-life,
> > > "production" gateways with tens
On Tue, Nov 01, 2011 at 09:47:35AM +0200, Gregory Edigarov wrote:
> On Tue, 1 Nov 2011 11:17:56 +0400
> ZZ Wave wrote:
>
> > What solution should be used for traffic shaping on real-life,
> > "production" gateways with tens and hundreds users? PF queues seem to
> > be too "userspace"-ish and CPU
On Tue, 1 Nov 2011 11:17:56 +0400
ZZ Wave wrote:
> What solution should be used for traffic shaping on real-life,
> "production" gateways with tens and hundreds users? PF queues seem to
> be too "userspace"-ish and CPU consuming.
Pardon?
What do you mean "userspace"-ish ?
--
With best regards
What solution should be used for traffic shaping on real-life, "production"
gateways with tens and hundreds users? PF queues seem to be too
"userspace"-ish and CPU consuming.
13 matches
Mail list logo