[Openstack] [Quantum] Why is 'subnet' introduced in Quantum API v2?

Salvatore Orlando sorlando at nicira.com
Fri Mar 15 10:06:53 UTC 2013


Hi Zhidong,

Please look inline for answers to your questions.

On 15 March 2013 08:05, Zd Yu <zdyu2000 at gmail.com> wrote:
>
> I am confused by the existence of both 'network' and 'subnet'. According to
> v2 API spec[1], subnet is introduced to support IPAM capabilities. Can
> anyone elaborate it to me? For example, what is the benefit of creating 1
> network and 2 subnets by comparison with creating 2 networks that implicitly
> has one and only one subnet for each?

Even if you're right that for most application you always have one
network and one subnet, conceptually they're different concepts, and
they have been kept separated to allow either L2-only networks (where
for instance, you might want to use your own IPAM system), or enable
multiple subnets on the same network, which is not an uncommon
scenario.

>
> I do know that the original network did not support IP overlapping before
> subnet is introduced. But it is just an implementation issue, right?
>

Folsom actually supported overlapping IPs, but they were switched off
by default because this created problem with nova's security groups
and metadata servers.
In the Grizzly release Quantum supports security groups natively, and
metadata can now work with overlapping IPs.
You just need to ensure that allow_overlapping_ips is set to True in
your quantum.conf.

> Thanks,
> Zhidong
>
> [1] https://wiki.openstack.org/wiki/Quantum/APIv2-specification
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~openstack
> Post to     : openstack at lists.launchpad.net
> Unsubscribe : https://launchpad.net/~openstack
> More help   : https://help.launchpad.net/ListHelp
>




More information about the Openstack mailing list