I wondered if there are any better approaches to performing network discoveries, especially on large networks, as it simply takes far too long, and discovers devices we simply don't care to monitor.
We have 2 /24 bit management networks for our L3 devices, but our L2 devices sit on our primary network which is /18 broken down into /24 or smaller subnets, so you can imagine this can take a long time to scan, and also picks up lots of devices we don't care about. Not all of our devices support snmp so we cant limit to discovering only these devices. Most of our access switches are on the primary network. I also don't want to have individual subnet groups, as there would be too many to manage.
Now, if we could only scan the management network and use CDP to discover connected switches, that would be ideal, but for some reason, CDP isn't offered, although I know SW is multi vendor, and CDP is cisco proprietary, surely it wouldn't be that difficult to provide such a useful feature to aid discovery?
We could then simply scan 2 /24 bit networks and if we could use CDP, it would discover all uplink devices!.. perfect!
Just wondered if I may be missing something, or if anyone had adopted any strategies for optimizing discoveries?