Publish Date: May 8, 2026
Executive Overview
Historically, setting up storage protection rules in a software-defined storage area network required deep engineering knowledge. Storage administrators had to manually build, assign, and modify precise Storage Policy-Based Management (SPBM) rules, adjusting parameters like “Failures to Tolerate” (FTT) and selecting specific RAID structures (RAID-1, RAID-5, or RAID-6) based on physical host constraints.
This post breaks down Auto-RAID in VMware vSAN for VCF 9.1. This feature replaces manual policy design with a fully system-managed, automated approach to data resilience. By enforcing a single, dynamic storage policy across the entire cluster, VCF 9.1 evaluates cluster sizes and disk topologies in real-time. It automatically shifts erasure coding levels up or down as hardware nodes are added or removed, removing the configuration complexity that traditionally led to sub-optimal space usage or accidental data exposure.
Features
The introduction of Auto-RAID moves storage availability from a user-defined manual entry into an intelligent, hypervisor-enforced runtime property.
- Universal Auto-RAID Storage Policy Template: Establishes a singular
vSAN ESA Auto RAID Policyon the vCenter Server that governs data resilience rules globally across all managed vSAN 9.1 clusters. - Dynamic Erasure Code Self-Adjustment: System logic that automatically detects changes in host counts and seamlessly switches between RAID types (e.g., automatically upgrading from RAID-1 to RAID-5 or RAID-6 as new physical server frames boot up).
- Automated Bootstrap Provisioning: Removes the legacy requirement for manual “Force Provisioning” overrides during initial setup, allowing new virtual machines to build smoothly on small or single-node configurations.
- Deprecation of Low-Level Storage Tweaks: Total elimination of legacy software constraints within the policy engine, including custom disk striping settings, manual flash read cache allocations, and manual compression rules.
- Intelligent Backend Re-Rebalancing Throttle: Integrated traffic policing that monitors real-time application workloads and applies background throttling to block migrations when a cluster changes size, preventing performance drops.
Benefits
By shifting data resilience management to system automation, VCF 9.1 yields direct operational and structural storage advantages.
- Consistent Cluster Overheads: Enforcing a single baseline policy ensures uniform capacity overheads across the entire datastore, matching the clear, predictable visibility of traditional hardware storage arrays.
- Maximized Flash Storage Spatial Efficiency: The platform automatically upgrades objects to highly efficient erasure coding (like RAID-5/6) as soon as the cluster reaches the required size, maximizing usable NVMe capacity without human intervention.
- Elimination of Configuration Drift and Risk: Automating RAID selection removes the risk of administrators accidentally deploying critical databases with inadequate protection rules or using space-wasting mirroring topologies on large clusters.
- Seamless Hardware Fleet Scaling: Data center technicians can drop new host nodes into a rack and attach them to a cluster without updating or auditing associated application storage policies.
Use Cases
The autonomous resilience features of Auto-RAID are built for rapidly expanding enterprise environments and service provider architectures.
- High-Growth Private Cloud Build-Outs: Bootstrapping small, three-node remote office clusters and scaling them up to comprehensive corporate data blocks over time, relying on the platform to automatically tune data protection layers along the way.
- Automated Elastic Container Nodes (VKS): Supporting scale-out container infrastructure where automated infrastructure blueprints are continuously spinning up new worker nodes that demand instant, perfectly optimized storage backing.
- Multi-Cluster Fleet Management: Allowing a lean platform engineering team to govern data protection rules across thousands of global hyperconverged nodes without manually managing separate, local storage policy configurations.
Alternatives
When determining how to configure data availability across software-defined arrays, enterprise architects evaluate this native update against other operational strategies.
- Manual Policy-Based Storage Design (Legacy vSAN / Point HCI Solutions): Maintaining traditional methods where administrators manually configure separate storage rules for different application types. While offering granular control, this leads to policy sprawl and leaves the infrastructure vulnerable to human error.
- Fixed Hardware-Controlled RAID Subsystems: Relying on physical, server-level RAID controller cards to aggregate disks before presenting blocks to the hypervisor. This path restricts visibility, blocks advanced software-defined optimizations, and creates single points of failure at the component level.
Alternative Perspective
While Auto-RAID significantly simplifies day-two data center administration, it removes a level of microscopic control that specific performance engineers rely on. For niche enterprise applications that demand intentional, sub-optimal configurations—such as isolating a highly intensive transaction log on a strict RAID-1 mirror within a massive 32-node cluster that the system defaults to RAID-6—Auto-RAID’s universal approach may restrict custom performance-tuning options. Administrators must selectively use the global “Apply Auto-RAID” override toggle to accommodate these exceptional, legacy workloads.
Final Thoughts
The introduction of Auto-RAID in VCF 9.1 highlights Broadcom’s broader engineering trend: replacing manual infrastructure tuning with declarative, system-managed automation. By encoding data protection rules directly into hypervisor intelligence, vSAN ESA removes the guesswork from storage resilience. In the high-density data center environment of 2026, making storage protection automatic ensures that as your private cloud scales, its safety and space efficiency scale right along with it.
Source
https://blogs.vmware.com/cloud-foundation/2026/05/08/auto-raid-in-vsan-for-vcf-9-1