Data transfer between collective members

In a Data Archive collective, you can use PI to PI to transfer time-series data from one collective member to another when the interface node cannot send that data directly. By design, n-way buffering distributes data directly from a data source to each Data Archive server in a collective, but system architecture and security restrictions can preclude this technique in some cases. Note, however, that deployments using PI to PI are not as robust as those using n-way buffering. For example, if the primary server becomes unavailable, the secondary server that receives data from PI to PI can only access historic data, not real-time data.

To use PI to PI to transfer data in a Data Archive collective, you must enable tag-attribute override parameters. Collectives require that each server have identical point definitions. If the primary server has points configured to receive data directly from the interface node, then each secondary server must have identical points configured to receive data directly from the interface node. Normally, PI to PI requires tags configured to receive data explicitly from PI to PI. However, after you enable tag-attribute override parameters, PI to PI can collect data for tags not configured explicitly for the PI to PI interface.

You can configure PI to PI on a Data Archive computer or you can configure PI to PI on separate computers. In the most basic configuration, you might configure PI to PI on a Data Archive computer to copy time-series data from the primary server to a single secondary server.

Configuring PI to PI on separate computers offers a more robust configuration. In a more complex configuration, you might configure PI to PI to copy time-series data from one or more servers in a control network to multiple servers in a business network. This configuration might include interface failover (that is, a redundant copy of the PI to PI interface) to ensure that a PI to PI interface is always running and copying data. This configuration must use N-way buffering to ensure that PI to PI copies identical data to all the servers in the business network. Finally, to ensure that source data is always available to the PI to PI interface, you configure source-server failover (a failover mechanism specific to the PI to PI interface).