1 - Will your replication system work in this situation?
Absolutely, the bottle neck will be the the VPN speed, which will not be a problem in your case.

2 - What will happen if we lose connectivity?
If the connection is lost then replication will begin again, and it only synch the data that was not transferred. This will happen because of the ECC check of the transmitted data. And as long as the replication is active and there are no error-messages, the copy is consistent.

This can easily be tested by downloading our Demo-CD from our website to replicate this issue to see how the data is consistent.

3 - If during a busy time it falls behind, will that leave site B unusable, or just a bit behind, but still OK?
It will still be OK, only the data that has not been transferred during the fail. Once connection is re-established then the synchronization of the data from the source will continue to destination as it should be. If site A fails, then you need to promote site B to become your source.

4 - What spec of server do you recommend to run this...
For what you have I would recommend a high speed dual core CPU (high cache count) and 4GB of RAM and a good controller that has the write back enabled. We tune our product on the Intel server from link below or follow the same specs.

http://www.open-e.com/data_storage_s...bserv=prodcomp

Write now the synch limitation is 4TB per volume, so it should be good enough for you. And if you have more than 4TB, you can create multiple volumes that each of them replicate 4TB. Please keep in mind that in the future releases we will fix this to replicate 16TB per volume.