Attached are modified pictures 5 and 9. In 5 I called the transport plane an example, and showed all resources terminating at the edge of the network. I also eliminated the diagram at the control plane level because i don't think it is needed - I think it may well become standard in the protocol doc, but it is not officially defined in NML and it seems pre-mature to include it in this. I think the diagram works ok, and I am fine with moving the connection capability to the connection section. In 9 I left the top picture in tact -- it seems fine to me. The bottom picture I made changes similar to what I did in 5. 5 doesn't show ports in the lower pic. I added the concept of ports at the end of links as well as at edge of switches, and show an SDP connection between ports. The important thing for me in this is that resources are in one net or another so they can be managed by an NSI/NRM. Connections between networks are physically null; topologically networks are interconnected with SDPs. I believe this is what we agreed at OGF in Chicago and before. Any comments? John