Hi, Thomas: Your proposed text improve a lot since it deal with both communication between two VN but also communication between one VN and one Non-VN, using general term "gateway" makes sense to me. Regarding my added text, what I am trying to say if we need to consider VM movement between two VNs (i.e.,VM mobility case), suppose, one VM is moved from VN1 to VN2, each VN has a gateway for forwarding traffic (Gateway1 for VN1, Gateway2 for VN2), When VM1 is moved to VN2 and is still communicating with VM3 in VN1, the Gateway1 which the VM1 is previously anchored to(VM3 is also anchored to Gateway1) is obviously changed to Gateway2. This is the case I claimed to be basic or simple case. In this case, we don't have more than one gateway in each VN for forwarding traffic, how do we deal with this use case? Shall we ignore it? Or Solution is obvious and it is not worth dealing with this case? It is not clear to me.
Regards! -Qin -----邮件原件----- 发件人: [email protected] [mailto:[email protected]] 代表 Thomas Narten 发送时间: 2013年4月18日 0:38 收件人: Qin Wu 抄送: Eric Gray; [email protected] 主题: Re: [nvo3] 答复: Comments on draft-ietf-nvo3-overlay-problem-statement Hi Qin. Thanks for the suggested text. It helps clarify the concerns you have. From your proposal, you seem to be raising the following: 1) We use the term "router" to refer to the device that relays traffic between devices connected to two different VNs, as well as for relaying traffic between devices A and B, where A is attached to a VN, while B is a conventional device not connected to a VN at all (e.g., it could be out on the Internet or on the DCN). You use the term "edge device" for this, but I think that is even less clear. I used the term "router" because it is acting much like a traditional router. It supports IP, and it it relays traffic between devices connected to different IP subnets. Maybe the term "gateway" would be better? 2) you've added some text, but I'm having difficulty seeing what it is trying to clarify. That said, how about I reword the first paragraph to say: OLD: Another problem area relates to the routing of traffic into and out of a virtual network. A virtual network may have two routers for traffic to/from other VNs or external to all VNs, and the optimal choice of router may depend on where the VM is located. The two routers may not be equally "close" to a given VM. The issue appears both when a VM is initially instantiated on a virtual network or when a VM migrates or is moved to a different location. After a migration, the VM's closest router for such traffic may change, i.e., the VM may get better service by switching to the "closer" router, and this may improve the utilization of network resources. New: Another problem area relates to the optimal forwarding of traffic between peers that are not connected to the same virtual network. Such forwarding happens when a host on a virtual network communicates with a host not on any virtual network (e.g., an Internet host) as well as when a host on a virtual network communicates with a host on a different virtual network. A virtual network may have two (or more) gateways for forwarding traffic onto and off of the virtual network and the optimal choice of which gateway to use may depend on the the set of available paths between the communicating peers. The set of available gateways may not be equally "close" to a given destination. The issue appears both when a VM is initially instantiated on a virtual network or when a VM migrates or is moved to a different location. After a migration, for instance, a VM's best-choice gateway for such traffic may change, i.e., the VM may get better service by switching to the "closer" gateway, and this may improve the utilization of network resources. Does that help? Thomas _______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3 _______________________________________________ nvo3 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nvo3
