On Mon, Oct 1, 2012 at 8:59 PM, Ryan Young <[email protected]> wrote:
> Hello all,
> I'm a newbie to openflow/ovs, and I have some questions regarding the
> hardware performance implementing Openflow.
> I got some experimental data below from some scientific
> reports(http://staff.science.uva.nl/~delaat/rp/2011-2012/p18/report.pdf, et
> al):
> (Varing with different platforms, mainly focuses on the commercial switches)
> The maximum number of  flows that can be installed in the flow table is
> around 100k;
> The time cost for installing single flow is about 1~6ms which means that
> intalling millions of flows may cost hundreds seconds.
> Does it means that recent openflow switches are still not applicable for the
> arduous cases of the large scale network, e.g the load balancing service for
> the web server cluster since the concurrent traffic may reach millions
> requests per second, and the hardware performance above is seriously
> unacceptable.

It entirely depends on the application, the design of the controller,
and the hardware.
_______________________________________________
discuss mailing list
[email protected]
http://openvswitch.org/mailman/listinfo/discuss

Reply via email to