Hi I am a beginner to Calico and trying to understand some of the basic architectural details.
Lets say one wants to use Calico for networking of docker containers running inside VMs (which may be Nova instances within an Openstack cloud and lets say that is using a VXLAN based overlay solution for its data plane). The first hop router is the linux router of the VM hosting the source container. In order to do optimal routing, one would require the first hop TOR device to terminate the Openstack VXLAN, map it to a network namespace for that exact openstack tenant, do the BGP routing, re-encapsulate into VXLAN and reinject into the Openstack Neutron data plane towards the Calico next hop. Alternately, one would need to accept "non-optimal" routing and have a Calico routing plane completely separate and overlaid on top of the IP fabric that transports the Openstack VXLAN data plane. Is this an accurate understanding of the packet forwarding options for the intermediate routing hops ? Or am I missing something. Thanks for any clarifications. Rgds, Sanjeev
_______________________________________________ calico-tech mailing list calico-tech@lists.projectcalico.org http://lists.projectcalico.org/mailman/listinfo/calico-tech_lists.projectcalico.org
