We have a main office network connected to 4 branches via internet connections. Each location has 2 internet connections, which include a T1-type connection, and a cable internet connection as a backup. Site-to-site VPN's exist using Sonicwall NSA240 at the main, and TZ100/200 at the branches. VPN's connect the branches to the main, and between the branches themselves as well (For interoffice dialing between the PBX's). Branch users leverage Terminal Services connections to the main office over the VPN, of which there are from 5 to 15 RDP sessions going on at any time in each branch.
We continually get complaints of RDP slowness from the branches, and are relativley sure it's not the servers in the main office. The sonicwalls use probing to manage the Failover and Load Balancing scenario, and sometimes bounce back and forth between primary and secondary internet circuits for no apparent reason.
1. What can we use to monitor the internet connections at the branches and the main office, so that we have a more accurate idea of where a problem lies between the branches, when slowness issues occur?
2. How could we re-engineer the WAN so that we dont continually have speed issues that we cannot resolve?
3. What are the probing settings on the Sonicwall (NSA and TZ series) that you might be able to recommend, that might provide a more accurate line failure detection?