|
On Sat, Apr 05, 2003 at 09:33:21PM -0500, John Reuning wrote:
> I'm glad someone asked about lvs-nat performance. I've read the
> performance tests and documentation. One test from 2000 or 2001 showed
> lvs-nat topping out at about 4000 connections per second; however, I'm
> more concerned with bandwidth. How much outbound data traffic can an
> lvs-nat director with dual e1000 adapters handle? Provided sufficient
> network equipment, file server performance, and real server performance,
> is 400-500 Mbps sustained, with peaks of 700-900 Mbps unreasonable?
I have done some testing which peaked out at about 700Mb/s because that
is how much traffic we could generate. So I would say that yes, you can
achive the level of throughput you are after. This was using e1000
adaptors and lvs-nat. Though some tweaking may be required.
Some things that you may want to consider are: Using a non SMP kernel -
there is actually more overhead in obtaining spinlocks than the
advantage you get from multile CPUs if you are only doing LVS. If you
only have one processor you should definately use a non-SMP kernel. You
may also want to consider using NAPI with the ethernet driver and if you
really want to use SMP then setting up affinity for the two NICS with
different processors is probably a good idea.
--
Horms
|
