We are running it on a VM with dual 2.8Ghz Xeon processors, 4GB RAM, 20GB C: drive for system and Netbackup client install, and 30GB F: drive for OpsCenter.
I would recommend looking at page 46 in the above technote for info on installing OpsCenter on a different drive. We've had problems when we change our OpsCenter logging levels and were filling up our C: drive on a daily basis with thousands of log files. Had to have Symantec enter an override in a config file that just overwrites the log file as it files up, which occurs every 3-4 seconds.
We're running OpsCenter on VMware virtual machine. It gives you management flexibility if you want to expand your machines's capacity. The virtual machines runs on 2 vCPU and 8GB vRAM. Giving more RAM would give a better performance as there is a OpsCenter has an embedded Sybase database.
For disk space, it's easy to expand for a virtual machine. 50GB is a good start.
Are you monitoring a single NBU Master? The OpsCenter Tuning and Best Practices recommends a stand alone server. I am looking to move our OpsCenter server to a new location so new server. I monitor approximately 10 sites/Masters with 2 of those sites in the 1k + client range.
Our system is just a single NBU master and a couple hundred clients. The only performance issues we've run into have been from the OpsCenter software itself (high CPU usage from the SQL database service). One of those was a bug in the OpsCenter software that caused timeouts. It would have happened on a physical server, too.
I can easily imagine that using OpsCenter to monitor 10 masters and thousands of clients would require a pretty beefy standalone server.
I'm a big proponent of following the info in the Best Practices guides, but this is one of the few we found that don't really apply to our smaller environment.