How IBM's Storage Controllers Keep Data Backups in the Right Order

A method for storage controllers to track and sequence data updates in a specific order, ensuring that remote backups remain consistent with the original data during a system failure.

What it covers

This patent describes a way for a storage controller to manage updates to data stored on a disk. When an application updates multiple records, the controller creates a circular queue to track these changes. Each update is linked to the previous one in a backward chain, and a counter keeps track of how many updates are pending. This allows a data mover to read the updates in the exact order they occurred, which is critical for sending them to a remote site for disaster recovery without data corruption.

What it doesn't cover

• —Does not cover data storage systems that lack a cache-based circular queue structure.
• —Does not cover methods of data transmission that do not require sequence-consistent ordering.
• —Does not cover the specific hardware architecture of the host processor itself.
• —Does not cover real-time data replication that occurs without a staging queue.

The clever bit

By using a backward-linked chain within a circular queue, the system can efficiently track the sequence of updates without needing to constantly re-sort or re-index the entire list of changes as they arrive.

Why it matters

This technology is a foundational piece of enterprise disaster recovery. It ensures that if a primary data center fails, the remote backup site has an identical, chronological history of data changes. This prevents 'data skew' where a backup might reflect a later state of one file but an earlier state of another, which would crash databases.

Real-world examples

• 1.Enterprise storage area networks (SAN)
• 2.IBM z/OS remote copy services
• 3.Disaster recovery replication software
• 4.High-availability database transaction logs