Streaming replication

When values of originating tier-1 tags are received from an IDAS or HCAL (AVEVA Application Server) and arrive at the tier-1 Historian as a time-ordered data stream directly, the Historian not only stores the data, but also forwards it to the Replication subsystem if replication is configured for those tags.

Then the Replication subsystem immediately streams that data to the tier-2 Historian for simple replication, or performs summary calculations and streams the resulting summaries. Likewise, if there are more tiers beyond tier 2, the Replication subsystem streams the data beyond tier 2 to the next-level tiers.

This happens equally efficiently for tag values of timestamps close to the current system time and for late data tags.

If any next-tier Historian becomes unavailable, the Replication subsystem continues to stream replicated data into the local store-and-forward path. When the connection is restored, all replicated data is sent as compressed snapshots to the next-tier Historian and incorporated into history.

Streaming replication is the fastest and most efficient way of data replication, but there are some scenarios where it cannot be used. In that case, another method called queued replication is applied.