In a distributed replication system, data is replicated at local and remote sites, so clients can continue to work regardless of what happens at the primary data source or over the WAN.
When a failure occurs at a remote site, clients can continue to use local copies of replicated data.
When a WAN failure occurs, clients can continue to use local replicated data.
When the local data server fails, clients can switch to replicated data at another site.
When WAN communications fail, Replication Servers at other sites store transactions in stable queues (disk storage) so that replicated tables at the unavailable site can be brought up to date when communications resume. When a replicated function is initiated in a source database, it is stored in stable queues until it can be delivered to the destination site.
