Deploying Dual-Active Storage with Virtualization and Application Clusters

Dual‑active storage is widely known, but deploying it together with virtualization and application clusters requires careful planning; this article discusses the necessary steps and considerations for such deployments.

Industries such as healthcare, securities trading, government, and manufacturing demand high availability for critical systems like Oracle, virtualization platforms, ERP, MS SQL, and HIS, making dual‑active solutions especially valuable.

As virtualization and cloud computing become prevalent, workloads often mix physical and virtual resources; OpenStack Ironic is highlighted as a technology that integrates these resources, illustrated with an Oracle database example.

When Oracle databases run in virtual environments, traditional primary‑backup disaster recovery (storage replication or GoldenGate) faces licensing costs and limited automation, leading many to prefer true dual‑active configurations.

Major storage vendors (Huawei HyperMetro, HDS GAD, Dell Live Volume, IBM HyperSwap, SVC ESC, Fujitsu Storage Cluster, HP PeerPersistence, EMC vPlex, NetApp MetroCluster, etc.) provide dual‑active features, and application‑level solutions like Oracle RAC with ASM or GPFS can also achieve active‑active storage.

Application‑level deployment options include physical clusters (Oracle RAC, DB2, Exchange, SAP HANA) that rely on OS or database clustering, and virtual deployments on VMware, Hyper‑V, or FusionSphere where the hypervisor’s HA and the application’s own HA are combined.

Two virtual deployment patterns are described: single‑VM deployment using the hypervisor’s HA for simplicity and proven reliability, and clustered‑VM deployment that adds nested clustering at the application layer but may encounter compatibility issues, so single‑VM is generally recommended.

The article then defines cloud storage, emphasizing self‑service, tenant isolation, SLA support, metering, and flexible billing models (Pay‑As‑You‑Go, Pay‑Per‑Use) as essential characteristics.

Examples of cloud storage implementations are provided, including enterprise drive‑by‑drive services, data‑service orchestration platforms (e.g., EMC ViPR, Huawei DJ), integration with cloud platforms via APIs, and hybrid‑cloud data movement solutions (e.g., IBM SmartCloud Storage Access, NetApp DataFabric, EMC CloudArray).