Worker connections are used to connect workers
Worker connections are used to link workers in different (or the same) clients to each other so that they can pass data between them. Each of the connected workers records the connection to the other using the Connections field.
Each worker takes a role in the connection, typically "owner" and "partner", in which the owner worker uses a role of "partner" to identify the partner, and the partner uses a role of "owner" to identify the owner. Each worker records its own role in the Role field. A worker can connect to multiple workers by using different roles for each.
Connected workers have the same process identifier, allowing the connections to route messages to the appropriate worker.
Local paths identify recipient folders
A worker in one client uses the Send process step type to send a process to another client. The other client creates a worker to interact with the first worker. As part of this, the sending worker can set a path, which is used by the recipient client to identify or create a folder in which to create the worker. Because the connection is itself identified by a path (the path of the connection's connection group), the path used to create the worker is termed the "local path" because it used locally at the recipient end.
Recipient local paths can be set in the outgoing connection's address book - see Group connections for details.
Connections data records potential and current connections
As well as setting the Connections field to record actual connections, workers use the connections data area to record connections, including:
- Assignments that will be used to make connections, before the connections are made.
- Details of requested connections, which use the Connection requests processing to create a connection by sending a request to somebody by email.
- Keeping track of multiple connections with the same logical role, such as multiple share connections for chat.
The connections data is maintained using Worker Connections Script, which documents the format of the data. The script is called automatically from within task processing or managed through the Connections step type.