Re: [ovs-discuss] SDN networkslower than physical network

[Guru Shetty]

Since this is the same host, your 25 Gbps physical network does not come to
picture. Neither does VXLAN as both the VMs are in the same host. I would
suggest trying the following experiments.

1. Multiple sessions of netperf/iperf.
2. Sometimes, your VM's virtual NIC is the bottleneck. So if you use
para-vritualized virtual NICs, you will likely get higher throughput.
3.  I would suggest adding 2 more VMs and running multiple sessions of
netperf/iperf to see your throughput increase.

On 5 July 2018 at 23:10, Manuel Sopena Ballesteros 
wrote:

> Dear OVS community,
>
>
>
> First of all, I am not sure if this is the right place for this question
> but I think it is worth to try.
>
>
>
> I installed openstack pike with ovs supporting the SDN virtual switches
> and had the impression that my SDN was not fast enough. Please see my
> configuration below:
>
>
>
> ·Physical compute hosts connected through a 25Gbps network.
>
> ·Each compute node has 64 cpus and 512 GB RAM.
>
>
>
> For some reason I could not get more than 10Gbps bandwidth of intra-subnet
> traffic (no routing) across 2 vms living in same physical host with VXLAN
> offload configured.
>
>
>
> I tested the network performance using iperf and iperf3 with multiple
> threads in parallel.
>
> No other vm was running in the system and the vms were only instantiated
> to run this network performance test so there were no shortage of resources
>
>
>
> Because the SDN traffic was within same compute node and same network (no
> routing involved) I am guessing that poor performance has something to do
> with OVS running in cpu but I have no clue to prove my theory. I am hoping
> someone in this email list would clarify if I am in the right direction or
> not and why.
>
>
>
> Thank you very much
>
>
>
> *Manuel Sopena Ballesteros *| Big data Engineer
> *Garvan Institute of Medical Research *
> The Kinghorn Cancer Centre, 370 Victoria Street, Darlinghurst, NSW 2010
> *T:* + 61 (0)2 9355 5760 | *F:* +61 (0)2 9295 8507 | *E:*
> manuel...@garvan.org.au
>
>
> NOTICE
> Please consider the environment before printing this email. This message
> and any attachments are intended for the addressee named and may contain
> legally privileged/confidential/copyright information. If you are not the
> intended recipient, you should not read, use, disclose, copy or distribute
> this communication. If you have received this message in error please
> notify us at once by return email and then delete both messages. We accept
> no liability for the distribution of viruses or similar in electronic
> communications. This notice should not be removed.
>
> ___
> discuss mailing list
> disc...@openvswitch.org
> https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
>
>
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Re: [ovs-discuss] SDN networkslower than physical network

[Manuel Sopena Ballesteros]

Dear OVS community,

First of all, I am not sure if this is the right place for this question but I 
think it is worth to try.

I installed openstack pike with ovs supporting the SDN virtual switches and had 
the impression that my SDN was not fast enough. Please see my configuration 
below:


*Physical compute hosts connected through a 25Gbps network.

*Each compute node has 64 cpus and 512 GB RAM.

For some reason I could not get more than 10Gbps bandwidth of intra-subnet 
traffic (no routing) across 2 vms living in same physical host with VXLAN 
offload configured.

I tested the network performance using iperf and iperf3 with multiple threads 
in parallel.
No other vm was running in the system and the vms were only instantiated to run 
this network performance test so there were no shortage of resources

Because the SDN traffic was within same compute node and same network (no 
routing involved) I am guessing that poor performance has something to do with 
OVS running in cpu but I have no clue to prove my theory. I am hoping someone 
in this email list would clarify if I am in the right direction or not and why.

Thank you very much

Manuel Sopena Ballesteros | Big data Engineer
Garvan Institute of Medical Research
The Kinghorn Cancer Centre, 370 Victoria Street, Darlinghurst, NSW 2010
T: + 61 (0)2 9355 5760 | F: +61 (0)2 9295 8507 | E: 
manuel...@garvan.org.au

NOTICE
Please consider the environment before printing this email. This message and 
any attachments are intended for the addressee named and may contain legally 
privileged/confidential/copyright information. If you are not the intended 
recipient, you should not read, use, disclose, copy or distribute this 
communication. If you have received this message in error please notify us at 
once by return email and then delete both messages. We accept no liability for 
the distribution of viruses or similar in electronic communications. This 
notice should not be removed.
___
discuss mailing list
disc...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss