Storage
Storage
The Storage domain tests how you match HPE Alletra, MSA, StoreOnce, and Zerto to workload and protection requirements. Expect block, file, backup, disaster recovery, and as-a-service consumption questions rather than speeds and feeds alone.
Core rules and concepts
- HPE Alletra 9000 is the mission-critical block platform: NVMe, active-active controllers, Peer Persistence for stretched clusters, and Recovery Manager Central for application-consistent snapshots.
- HPE Alletra 6000 is Nimble-based all-NVMe storage with CASL, scale-up architecture, and AIOps-driven predictive analytics. Alletra 5000 is the hybrid-flash Nimble-based option for cost-sensitive mixed workloads.
- HPE Alletra Storage MP is a disaggregated, multi-protocol architecture for block and file, delivered through HPE GreenLake for Block Storage and File Storage. HPE MSA is the entry-level array.
- HPE StoreOnce provides deduplicating backup with Catalyst, replication, and Cloud Bank to object storage. HPE Zerto provides journal-based continuous data protection with recovery point objectives in seconds for virtualized workloads.
- GreenLake storage services are consumption models; the underlying array family still determines capability and placement.
Worked example
A hospital needs Tier-0 electronic health records with near-zero downtime across two data centers, plus backup retention and VM-level disaster recovery. Place the EHR on Alletra 9000 with Peer Persistence across the two sites and a third-site witness for quorum. Use Recovery Manager Central for application-consistent snapshots. Use StoreOnce for deduplicated backup retention and Catalyst replication. Use Zerto for VM-level continuous protection and recovery orchestration. The reasoning is workload separation: active-active block storage for Tier 0, dedupe backup for retention, and journal-based CDP for virtualized DR.
Traps that cost marks
Alletra 9000 is not the Nimble-based 6000; Peer Persistence is not asynchronous replication. Backup is not disaster recovery. StoreOnce deduplication ratios are workload-dependent, not guaranteed. Zerto is not a backup repository. GreenLake is a consumption and operations model, not a replacement for array selection.