The factor here is really Ops Mgr - not nworks. We cannot force the entire topology view through in one discovery, or we would overload the Ops Mgr Root Management Server - and so we would impact the entire Ops Mgr environment (in larger enterprise deployments, anyway).
This is why the topology discovery is cascaded - we discover it in several stages initially, first Clusters, then Hosts, then VMs etc. So that does mean it takes a few collection cycles to build up....but once it is built, it will be on-demand dynamically updated, when you add hosts, or move VMs, etc.
5.7 has various improvements in topology scalability, performance, and overall stability - I know you're on the Beta program, so you should be seeing that in your lab right now!