According to your description, you may consider Peer-to-Peer replication to synchronize data between multiple servers to keep the relatively in real time data on each node. It might be a big deal to collect changes from multiple servers and apply them manually
by taking use of CDC or timestamp. Please pay attention to this article about
Peer-to-Peer Replication in SQL Server 2008.