Hi,
I think it will be very hard for any one in the communities to answer this for you.
Below are some of the reasons why
1) this depends on how many devices are being monitored
2) how often each qos is being monitored
3) how many alarms are being generated and how often
4) the underlying hardware especially disk IO for the secondary hubs and primary
5) the network bandwidth between locations.
( generally speaking links between hubs that are under 1MBS will usually be an issue on busy networks if they are dediated
if they are not dedicated then this would nee to be much higher in most cases)
( high latency network connections will cause you issues as well as if you connections are using a vpn layer)
I would suggest engaging your pre-sales engineer as they do have some excel spread sheets that can help you plug in
the information above to figure out the bandwidth and database needs based on what you are doing in your environment.
I do not currently have copies but if some one else has them maybe they can attach them for you.