Hi Sergiy ,
Thanks for your reply . The case of splitting the identity ranges , such that one server will have positive identity ranges and the other will have negative identity ranges will work as long as we only have 2 PEERS ( servers ) in the replication topology . This will break if we want to have 3 or more peers ( servers ) in the replication topology.
Is there any way we can do what SQL Server replication is doing ?
Thanks for your help .