network names are definitely not unique in neutron and I don't think it would be tenable to make them that way. Ignoring the backward-compatibility, it would make the acceptable name field values for a given tenant completely unpredictable because they can't see all of the names that are in use.
Even if we scoped uniqueness to (network tenant_id, network name), it would still result in cases where just the name collides when you have networks shared to a tenant either globally or via an RBAC entry.
I think the Nova API is going to have to change to key based on network IDs instead of names or one of the other solutions.
network names are definitely not unique in neutron and I don't think it would be tenable to make them that way. Ignoring the backward- compatibility, it would make the acceptable name field values for a given tenant completely unpredictable because they can't see all of the names that are in use.
Even if we scoped uniqueness to (network tenant_id, network name), it would still result in cases where just the name collides when you have networks shared to a tenant either globally or via an RBAC entry.
I think the Nova API is going to have to change to key based on network IDs instead of names or one of the other solutions.