I am using OVS to build logically seperated Layer 2 networks over the top of a Layer 3 network using GRE tunnels to create a fully connected mesh for each network.
This works perfectly and is very scalable. I have had no issues creating pretty obscene numbers of GRE tunnels on a single host. With this setup the guest VM networks are completely abstracted from the physical underlying network (which in this case doesn't even support Layer 2 Ethernet!) Due to the nature of the GRE tunnels being Layer 3, optimal routing is preserved and the implementation is efficient hinging on OVS being well behaved (which is seems to be) Joseph. On 8 September 2011 16:04, Justin Pettit <[email protected]> wrote: > The plan is to support both models. > > --Justin > > > On Sep 7, 2011, at 10:54 PM, Holger Winkelmann wrote: > > > Sounds good to me too. To clarify, you 'just' Plan to support the > encapsulation format and control will be still the OF controller ? > > > > -- > > > > Holger Winkelmann > > Travelping GmbH > > +49-171-5594745 > > > > On 08.09.2011, at 07:36, Justin Pettit <[email protected]> wrote: > > > >> Right, I agree. I was just referring to the ability to use its UDP > tunneling format to get through devices that don't like GRE. We're also > planning to support a mode where VXLAN is a virtual port, like GRE, and > traffic that is sent out it will be encapsulated (and decapsulated on > receive). > >> > >> --Justin > >> > >> > >> On Sep 7, 2011, at 10:03 PM, Nicky Fatr wrote: > >> > >>> Yes, VXLAN tunnel header is a good proposal, but for control plane > >>> there is serve limitation: it depend on physical network multicast for > >>> MAC learning. In OVS, central ovsdb controlled MAC address propagation > >>> is a better choice. > >>> > >>> On Wed, Sep 7, 2011 at 9:09 AM, Justin Pettit <[email protected]> > wrote: > >>>> On Sep 7, 2011, at 9:01 AM, Nicky Fatr wrote: > >>>> > >>>>> I don't think that TRILL/802.1AQ L2 over L2 is a good option for > large > >>>>> scale deployment. L2 over L3 instead is more scalable, eliminating > >>>>> comlexity of physical network. > >>>>> > >>>>> maybe we can expect L2 over UDP in some future release, for UDP is > >>>>> more friendly than GRE in some networking configuration. > >>>> > >>>> You can already do L2-over-L3 with CAPWAP. It doesn't support a > configurable context identifier (key), but a patch has been provided by > Valient Gough and Simon Horman that adds it. We're also looking at > supporting VXLAN, which was recently announced: > >>>> > >>>> http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-00 > >>>> > >>>> --Justin > >>>> > >>>> > >>>> > >> > >> _______________________________________________ > >> discuss mailing list > >> [email protected] > >> http://openvswitch.org/mailman/listinfo/discuss > > _______________________________________________ > discuss mailing list > [email protected] > http://openvswitch.org/mailman/listinfo/discuss > -- * Founder | Director | VP Research Orion Virtualisation Solutions* | www.orionvm.com.au | Phone: 1300 56 99 52 | Mobile: 0428 754 846
_______________________________________________ discuss mailing list [email protected] http://openvswitch.org/mailman/listinfo/discuss
