Right now, for one instance of the connector, all configured data sources will send data to all configured PI Data Archives and PI AF Servers. It is all or nothing. Due to the fact that you cannot have multiple instances of the same connector on any given node at a time, if you want this kind of granular control you would need an entirely new node and another connector. In the past, it was trivial to have multiple interfaces instances on the same node to send different data to different PI Data Archives; connectors should allow this capability.