Replication Server keeps track of other components in a replication system using connections that identify primary and replicate databases and routes that identify other Replication Servers.
Since Replication Server was originally designed for Adaptive Server Enterprise database replication, the definition of a connection in Replication Server follows the Sybase standard of <server name>.<database name>. For example, a Replication Server connection to an Adaptive Server named ASE1 and database PUBS is named ASE1.PUBS.
To connect to a primary non-ASE data server, Replication Server allows a connection from a Replication Agent on behalf of the non-ASE primary database. For a replicate database, Replication Server connects to an ECDA database gateway, which in turn connects to the non-ASE replicate data server. For Oracle, Replication Server can also connect directly to the replicate data server using ExpressConnect. Since Replication Agents, ECDA gateways, and ExpressConnect are not data servers, the Replication Server connection properties for those components may have different meanings than they do for a database server connection.
A single Replication Server connection can support data flow in either one or two directions. Data flows in through a Replication Server connection by way of the Replication Agent user thread. Data flows out through a Replication Server connection by way of the Data Server Interface (DSI) thread. Each Replication Server connection can support either outbound data flow only (through the DSI thread), or both inbound and outbound data flow (through the Replication Agent User and DSI threads).