Dynamic nsrmmd

I would urge you to actually take a Networker basics course or at least read the actual NW manuals that are available, as dynamic nsrmmd's are explained (a bit) in there. Or even look in the Help when in NW NMC console...

"If Dynamic nsrmmds is selected, then the NetWorker adds devices only on demand. Dynamic mode can improve bandwidth and performance, but the firewall ports may be left unattended by running processes. If Dynamic nsrmmds is set to No, then the NetWorker starts all available nsrmmd processes. Static mode along with the "mmds for disabled devices" option provides greater control on the nsrmmd firewall ports. The ports are attended by running processes."

If the NW client version and the used DDOS version are in supported ranges of each other, then one can have source side deduplication occurring (what NW calls "client direct"), which means that clients should be able to reach a DD system directly. If they can't then the configured Networker storage node (NW SN) will be used to send data towards and dedupe will be performed by the NW SN, which means the backup is full blown send to the NW SN before dedupe is being done, causing a higher load on the NW SN.

But as a DD system is very versatile, one can also put backup data on cifs or nfs shares, which would mean target side deduplication. Or using it as a virtual taper library (VTL) on a SAN. However if possible, ddboost protocol used by the NW clients to talk to the DD, is the way to go. At least for us it is. Haven't looked back after we got rid of all physical and virtual tape altogether...

But that means more or less buying into the Dell ecosystem of both backup software and backup targets. We went all-in really.