The DSP package is intended to allow applications to run in high reliability mode, and support an unlimited number of Thin Client connections to either of the servers running those redundant applications. Applications that are running on two servers need only have Thin Clients licensed on one of the two servers. When the servers run the same networked application, the two servers will pool their Thin Clients, allowing that total number of concurrent Thin Client connections from the network of servers. To enable this pooling of Thin Clients, both servers' IP addresses must be configured in the Internet Client/Server setup (Server tab) on each server.
For your specific situation, assuming one of the server licenses will be installed at the WTP and the other at the WWTP, the key will be in setting each of the applications up as redundant on both servers. The WTP application will have its local server as the Primary and the WWTP server as the backup. The WWTP will have its local server as the Primary and the WTP server as the backup. The loading between the servers will define the loading on the Internet connection. If this is a reasonably high bandwidth and generally reliable connection, the server to server loads shouldn't be significant, however, a couple of pointers...
Realtime data exchange between the servers. This can be very high if no deadbanding is used on Analog tags and values are changing constantly in the PLC. Start with deadbands and adjust these down if needed.
Historical data exchange between the servers. This can be similarly high if no deadbanding is used. Use deadbanding here as well.
If you have very low bandwidth on the Internet connection, you could set up the local server at each location as the only Historian server. While this would limit server-to-server traffic, you'd lose continuous data redundancy.
The DSP package is intended to allow applications to run in high reliability mode, and support an unlimited number of Thin Client connections to either of the servers running those redundant applications. Applications that are running on two servers need only have Thin Clients licensed on one of the two servers. When the servers run the same networked application, the two servers will pool their Thin Clients, allowing that total number of concurrent Thin Client connections from the network of servers. To enable this pooling of Thin Clients, both servers' IP addresses must be configured in the Internet Client/Server setup (Server tab) on each server.
For your specific situation, assuming one of the server licenses will be installed at the WTP and the other at the WWTP, the key will be in setting each of the applications up as redundant on both servers. The WTP application will have its local server as the Primary and the WWTP server as the backup. The WWTP will have its local server as the Primary and the WTP server as the backup. The loading between the servers will define the loading on the Internet connection. If this is a reasonably high bandwidth and generally reliable connection, the server to server loads shouldn't be significant, however, a couple of pointers...
1. Realtime data exchange between the servers. This can be very high if no deadbanding is used on Analog tags and values are changing constantly in the PLC. Start with deadbands and adjust these down if needed.
2. Historical data exchange between the servers. This can be similarly high if no deadbanding is used. Use deadbanding here as well.
3. If you have very low bandwidth on the Internet connection, you could set up the local server at each location as the only Historian server. While this would limit server-to-server traffic, you'd lose continuous data redundancy.
