Understanding HPE 3PAR Storage Federation, Peer Motion, Online Import, and Peer Persistence

HPE 3PAR storage offers a powerful feature called Storage Federation , which provides three main capabilities: Peer Motion , Online Import , and Peer Persistence . Peer Persistence implements an asymmetric active‑passive dual‑active solution and its architecture is illustrated with diagrams.

Peer Motion enables free movement of resources across a federation of devices, supporting migration of all LUNs mapped to a host or all hosts mapped to a LUN. It allows online, non‑disruptive migration between 3PAR StoreServ systems, with bidirectional data flow and no need for extra migration hardware.

Online Import builds on Peer Motion to migrate active data without interruption from heterogeneous storage arrays such as EMC VMAX, VNX, CLARiiON, HDS NSC/USP/VSP, IBM XIV, etc., to a new 3PAR StoreServ, supporting consistency‑group migration.

Beyond Peer Persistence, 3PAR provides other persistent features: Persistent Cache (mirroring cache on controller failure), Persistent Ports (redirecting I/O during controller upgrades without host multipathing), and Persistent Checksum (SCSI T10 PI‑based end‑to‑end data integrity for flash).

Peer Persistence creates an active‑passive dual‑active cluster that maintains business continuity during server failures. Compared with IBM SVC and EMC VPLEX, it embeds the functionality in the storage itself, avoiding extra virtualization gateways, though it requires identical storage models.

Using an Oracle RAC example, the article shows how storage is connected via FC, with Peer Persistence presenting a single logical volume to the RAC cluster. It explains why Peer Persistence is not true active‑active: I/O is first directed to the primary array (Active/Optimized) and then replicated to the secondary (Active/Unoptimized) via ALUA, and failover swaps these roles.

The discussion then compares active‑active and active‑passive approaches, covering switch speed, automatic failover latency due to arbitration, load‑balancing trade‑offs, and the impact of lock mechanisms on performance.

Other vendors' solutions are also described: Dell SC series Live Volume (active‑passive using synchronous replication and multipathing) and Fujitsu ENTERNUS DX S3/S4 Storage Cluster (active‑passive using Transparent Failover Volumes, CA ports sharing WWPN/WWNN, and a cluster controller). The article outlines the failover steps for the Storage Cluster, noting a typical 3‑second switchover.

Overall, the piece provides a detailed technical overview of 3PAR's storage federation capabilities, their operational characteristics, and how they compare with similar high‑availability storage solutions.