Robert, -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Robert Raszuk Sent: Thursday, June 28, 2012 11:09 AM To: John E Drake Cc: [email protected]; [email protected]; [email protected] Subject: Re: [nvo3] call for adoption: draft-narten-nvo3-overlay-problem-statement-02
> Robert, > > I expect the CE-PE interface, aka VPN service interface, will be > implemented between a multiplicity of devices. However, my point was > a different one. And my point is that I do not really see a clear need to run any PE-CE dynamic routing protocol if we assume that PE is running on the end host. I am also not actually sure if this is in scope of NVO3 to worry about PE-CE dynamic routing protocol as CEs would most likely be VMs regardless where the PE is and therefor to simplify the architecture they may just default to the PE. Such default would be provisioned by the orchestration system just like the IP address of the VM interface so IMHO out of scope of this work. [[LY]] I pretty agree on the view. VM was typically configured with its IP address, GW IP address, DNS address. Here seems no longer need to configure GW IP address in an VN, default to the NVE. I think this should be within the scope of nv03. Lucy Thx, R. > Viz, L3/L2 VPNs do not require BGP as the CE-PE interface. So, > rather than BGP being a 'non-starter', it is invisible. L3VPNs also > support OSPF v2 & V3, as well as others I have probably forgotten. > E-VPNs support STP, LAG, LLDP, etc. There are also Pedro's proposal > and the IEEE's VDP. > > Thanks, > > John _______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3 _______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
