Hi, here is another attempt to define what a virtual network in the sense of network virtualization is: A virtual network (VNet) is a set of (virtual) nodes directly connected by (virtual) links and realized on top of a set of underlying physical resources, the "substrate". This essentially represents the view of a VNet owner/user, so he sees virtual nodes that are directly connected to other virtual nodes, i.e., a bare link layer topology. Protocols running _inside_ the virtual topology are typically layer 3 protocols, therefore we speak of Network Virtualization (virtualization at the network layer, and virtual nodes are thus virtual routers) Some features of VNets: - A virtual network resource (node or link) appears to a user of that resource as if he is the (exclusive) owner of that resource. (=> isolation) - There should be no assumptions about the particular network protocols or architectures running inside the VNet, i.e., it is not necessarily IP. - One may use various substrate techniques (even at different layers) to create virtual links, e.g., WDM, IP Tunnels, MPLS, Ethernet VLANs, or even TCP tunnels. The substrate may be composed of heterogeneous technologies. The main interest of the VNRG is probably to consider IP(v4/v6) related technologies as common substrate technology. - Partitioning or aggregation of substrate resources is possible to create virtual resources. - A host that wants to connect to a VNet needs a running instance of the protocol(s) that are required to communicate with the VNet. It needs a substrate path (i.e., the virtual last mile link) in order to attach to a virtual (access) node. Question: does the host "belong" to the VNet topology or not? IMHO, it is the same situation as in the Internet today: hosts are part of the network but don't belong to an ISPs infrastructure, i.e., they are attached to the access routers/switches etc. Another interesting question is, however, how the end of a virtual link is represented inside the virtual node, e.g., it could appear as an "ordinary Ethernet" device (i.e., virtual device) or as any other L2 device. Regards, Roland
Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.