Dell and Microsoft have taken a notable step toward making enterprise-grade unstructured storage a first-class citizen inside Azure by opening the public preview of a fully managed, Azure-native PowerScale offering that brings Dell’s OneFS to the Azure portal and marketplace experience.
Background / Overview
The new public preview expands the existing choices for running Dell PowerScale in public cloud environments. Until recently, enterprises that wanted PowerScale in Azure had to deploy a customer-managed edition — provisioning virtual machines, networking, and storage inside their own Azure subscriptions and managing OneFS themselves. The public preview introduces a second flavor: a Dell-managed, Azure-native ISV service that runs as a transactable resource through the Azure Portal and Azure Marketplace. This change converts PowerScale from an application you deploy into a Filesystem-as-a-Service-style offering that can be provisioned, monitored, and billed through Azure tooling and purchasing channels.
Microsoft’s Azure-native integration and Dell’s APEX positioning are central to this move. The offering promises to preserve OneFS capabilities — the shared global namespace, multi-protocol access, data reduction features, CloudPools tiering, and replication tools — while shifting the infrastructure management burden to Dell for customers who choose the managed variant. The result is intended to be an easier on-ramp for organizations looking to run data-intensive workloads in Azure without re-architecting storage or relinquishing enterprise storage features.
What is Azure Native Dell PowerScale?
Two deployment models: choice and control
The product family now exposes two clear pathways:
- Customer-managed Dell PowerScale for Azure — customers deploy the PowerScale OneFS stack themselves inside their Azure subscription. They control compute, networking, storage choices, and lifecycle operations. This model provides the most control for customization, integration, and compliance but also the greatest operational overhead.
- Dell-managed (Azure Native) Dell PowerScale / APEX File Storage for Azure — Dell provisions and operates the underlying infrastructure and OneFS platform, while customers consume the filesystem from their Azure environment. The service is offered as an Azure Native ISV integration and is transactable through Azure Marketplace so customers can use existing Azure contracts and consumption commitments.
Azure-native ISV integration: what changes for admins
Being implemented as an
Azure Native ISV service means several practical changes for IT teams:
- The service is visible and consumable via the Azure Portal, and it can be managed with Azure tooling such as the Portal, Azure CLI, or PowerShell, matching administrators’ existing workflows.
- Procurement flows through Azure Marketplace, enabling use of Microsoft Enterprise Agreements, marketplace private offers, and Azure consumption commitments.
- The offering integrates into Azure resource and subscription models, including role-based access control (RBAC), metrics collection in Azure Monitor, and Azure billing.
These integrations are designed to lower friction for organizations that want enterprise file services tightly integrated with Azure compute and AI services.
Technical capabilities — what OneFS brings to Azure
OneFS fundamentals preserved
PowerScale’s OneFS is the enterprise-grade distributed filesystem at the heart of the offering. The key OneFS capabilities retained in cloud editions include:
- Single global namespace and a single filesystem spanning nodes that simplifies application integration and capacity management.
- Multi-protocol access: NFS (v3/v4), SMB3, HDFS and S3 access (OneFS supports S3 as a first-class protocol), enabling a wide range of legacy and cloud-native apps to access the same data set.
- Data services: snapshots, SmartDedupe (deduplication), inline compression, SmartQuotas, SmartConnect (client load balancing), SmartLock (WORM/immutability) and snapshot-based recovery.
- CloudPools: policy-driven tiering that archives colder files to object storage (Azure Blob Storage) while keeping metadata and a smartlink pointer on the primary filesystem.
- SyncIQ replication: asynchronous replication for disaster recovery and hybrid workflows between clusters (on-prem and cloud).
These are the capabilities enterprises expect from PowerScale on-premises — the cloud editions aim to preserve them while adapting management and operational models for Azure.
Scalability, performance, and stated limits
Dell’s public materials for the customer-managed Azure edition indicate clustering scale limits and cluster capacities that are meaningful planning inputs:
- Customer-managed clusters in Azure have been presented with scale guidance of up to 18 nodes and 5.6 PiB of usable cluster capacity in Azure for that edition (PiB — pebibyte — is the binary measure often used in storage specs). This maps to multiple petabytes of usable capacity and is aimed at large dataset workloads typical in AI, analytics, HPC, and EDA.
It is important to note that published capacity figures vary by edition and deployment architecture. Dell’s on-prem hardware families show far greater theoretical maximums (many tens of petabytes) when using physical PowerScale chassis; cloud editions are constrained by cloud VM types and configuration choices. The managed Dell-native edition’s maximum filesystem capacity and node counts are subject to the managed service configuration and may differ; when exact capacity needs are critical, customers should request exact limits from Dell during planning.
Flagged claim: some third‑party coverage has cited an
8.4 petabyte single-namespace figure for the managed Azure edition. That specific number could not be confirmed in Dell or Microsoft documentation available at the time of writing and should be treated as
unverified until Dell or Microsoft publish explicit capacity limits for the Dell-managed, Azure-native edition.
Data mobility, hybrid features, and AI readiness
The PowerScale architecture for cloud is explicitly aimed at data-heavy, throughput-sensitive workloads:
- Hybrid optimization — CloudPools tiering, SyncIQ replication, and object connectors enable replication and staged data movement between on-prem PowerScale clusters and the Azure-based PowerScale instances. This supports use cases such as cloud bursting for compute-intensive workloads, hybrid disaster recovery, and phased migrations.
- AI/ML and HPC support — the scale-out architecture and parallel file access model are aimed at providing large datasets and high throughput for training models, feature engineering, and HPC workloads. Dell and Microsoft position the Azure-native service to work tightly with Azure AI services and GPU-based compute instances.
- File-to-object integration — OneFS’s S3 support and CloudPools allow applications expecting object storage to use PowerScale datasets, while policies push cold data to Azure Blob for cost efficiency.
Operational model: what changes when Dell manages the stack
Responsibilities and support
Under the
customer-managed model, the customer handles:
- Provisioning of the virtual machines or Azure marketplace deployment template resources.
- Network design, subnet and NSG configuration, plugin drivers, and performance tuning.
- OS, storage and OneFS lifecycle management (patching, upgrades).
- Monitoring, backups, and integration into the customer’s SLAs/DR plans.
Under the
Dell-managed Azure-native model, Dell assumes the responsibilities for:
- Underlying infrastructure provisioning, maintenance, platform monitoring, and upgrades.
- Providing a support boundary where Dell is the primary contact for hardware/software platform issues.
- Ensuring the OneFS software stack remains current and stable as a managed platform.
This model reduces operational effort for customers but introduces a different support relationship and potential constraints on deep customization of the underlying infrastructure.
Integration with Azure subscriptions and billing
Because the managed service is transactable through Azure Marketplace:
- Customers can often apply existing Azure enterprise discounts or consumption commitments.
- Billing and chargeback flow through Azure subscriptions and the existing Microsoft billing pipeline.
- Private Marketplace offers and negotiated terms may be supported, but the exact billing mechanics and whether certain discounts apply will depend on the Azure Marketplace offer model and any negotiated EA/MCA terms.
Benefits — why this matters for enterprise IT
- Fewer ops headaches: Dell manages the infrastructure lifecycle, enabling customers to focus on applications and data rather than low-level cloud resource plumbing.
- Preserves enterprise features: OneFS data services (dedupe, snapshots, replication, and multi-protocol access) bring enterprise-grade data management to Azure.
- Hybrid continuity: SyncIQ and CloudPools provide mechanisms to orchestrate hybrid data estates, supporting migration, DR, and cloud bursting.
- Native Azure consumption: Marketplace transactable resources mean simpler procurement, billing consolidation, and potential use of committed Azure spend.
- Built for data-intensive workloads: The platform targets AI/ML, HPC, EDA, and analytics workloads where throughput, parallel access, and large datasets are key.
Risks, caveats, and what to watch
No technology is a silver bullet. The new Azure Native Dell PowerScale service introduces trade-offs and areas that demand careful evaluation.
1. Capacity and scale confusion — verify hard numbers
Published scale figures differ between materials for customer-managed and Dell-managed editions. For planning, treat scale and capacity numbers as
contractual variables until you receive explicit published limits from Dell or Azure for the managed offering. Do not rely on third‑party articles for hard capacity guarantees.
2. Potential vendor lock-in and data gravity
Moving large volumes into a managed appliance in Dell’s tenant (the Dell-managed model) can increase
data gravity and make future migrations or multi-cloud strategies more complex and costly. Evaluate the exit strategy including data export speeds, egress costs, and supported data mobility tools.
3. Cost model and egress/ingress economics
Cloud economics matter: storage, network egress, snapshot retention, and cross-region replication can multiply costs. When workloads access data from Azure compute across availability zones or regions, data transfer pricing and performance should be modeled and validated with proof-of-concept tests.
4. Latency and performance predictability
Although OneFS is engineered for high performance, cloud-based file systems are subject to variability in VM-to-storage network paths, VM host performance variability, and noisy-neighbor effects. For performance-sensitive AI/HPC workloads, test end-to-end throughput, concurrency, and latency under realistic loads.
5. Security, compliance, and control surface
A managed service introduces third-party control over part of your infrastructure stack. Ensure that the support model and service agreement meet your compliance and audit requirements. Ask about:
- Data residency guarantees and region-level placement controls.
- Encryption at rest and in transit, and how keys are managed.
- Access logs, auditability, and integration with your SIEM and Azure AD.
- RTO/RPO commitments for production workloads, and the incident response process from Dell.
6. Network design and delegated subnets
Azure Native integrations require
subnet delegation to the managed service resource provider. That requires careful network planning, peering, routing, and security group design to ensure the service meets organizational network architecture and security policies.
Practical checklist for IT teams evaluating the preview
- Confirm the exact service edition you plan to use (customer-managed vs Dell-managed) and map responsibilities for patching, monitoring, and support.
- Request and document published capacity numbers and node limits for the managed edition; get them in writing where capacity is a gating factor.
- Validate region availability for the preview and planned GA; check data residency implications and whether the regions you need are supported.
- Perform proof-of-concept testing for target workloads (AI training, HPC file I/O, large-scale analytics) to validate throughput and concurrency.
- Model costs including storage, snapshots, replication, and network egress. Include Azure Marketplace billing implications for chargeback and committed spend.
- Confirm backup/DR and replication strategy — see how SyncIQ and CloudPools interplay with your on-prem backup systems.
- Review security and compliance artifacts: encryption, key management, access logs, attestations, and third-party audits.
- Plan network architecture: delegated subnet, NSG rules, peering, private endpoints, and any ExpressRoute or VPN capacity required for hybrid operations.
- Understand support SLAs and escalation paths: who to contact for infrastructure vs. application issues and SLA targets for incident response.
- Verify whether the managed edition supports the advanced OneFS features you rely on (SmartDedupe, SmartLock retention policies, S3 protocol behavior).
How this fits into broader cloud storage strategy
The managed PowerScale experience in Azure is one more example of a broader market trend: making enterprise-grade storage services available natively in hyperscaler portals. For organizations with large unstructured data estates, the managed model offers the promise of rapid scale and operational simplicity while preserving feature parity with on-prem appliances. For others, the customer-managed model preserves control and customizability.
Two strategic considerations stand out:
- For organizations heavily invested in Azure AI services and looking to avoid costly re-platforming or data migrations, an Azure-native OneFS path reduces friction for bringing large datasets into cloud-based model training and analytics.
- For teams that require tight control for security, compliance, or specialized networking, the customer-managed deployment will remain the preferred path until managed offerings can show strong controls and flexible placement guarantees.
Vendor and marketplace implications
Delivering a managed PowerScale through the Azure Marketplace has several implications:
- Procurement simplicity: customers can buy through Microsoft channels, potentially simplifying contracting and enabling use of existing Azure discounts or enterprise agreements.
- Partner/reseller dynamics: organizations will need to map internal purchasing workflows (who signs off on Marketplace purchases) and whether private offers or reserved instances are supported for managed services.
- Support consolidation: Dell will be the primary platform operator in the managed model, which centralizes responsibility but requires confidence in Dell’s operational SLAs and responsiveness.
Final assessment and recommendations
The public preview of an Azure-native, Dell‑managed PowerScale offering represents a logical and valuable extension of enterprise storage into Azure for organizations that need a feature-rich, multi-protocol, and scalable file system without adopting a cloud-native re-architecture.
Key takeaways:
- Positive: The offering promises to deliver OneFS’s enterprise capabilities in an Azure-native form, integrate with Azure purchasing and tools, and simplify operations for large-scale file-based workloads.
- Caveats: Capacity numbers reported in secondary coverage are inconsistent; primary documentation and Dell/Microsoft materials should be used for planning. Cost, latency, compliance, and data-mobility considerations remain critical and require proof-of-concept validation before production adoption.
- Actionable steps: Teams should run a scoped pilot, verify required region availability, gather published capacity limits from the provider, test workload performance, and model costs — particularly network egress and long‑term tiering charges.
Unverified claims flagged: some coverage cites an 8.4‑petabyte single-namespace figure for the managed edition; this number was not verifiable against Dell or Microsoft product pages at the time of writing. Dell’s published customer-managed Azure guidance references up to
5.6 PiB per cluster for that edition. Customers should rely on official Dell/Microsoft documentation and marketplace offer details for contractual capacity and availability statements.
The Azure-native PowerScale public preview is an important evolution for enterprises needing scale-out file services within Azure. By combining OneFS data services with Azure-native procurement and management, the offering aims to make large-scale unstructured data usable for cloud-first AI and analytics workflows — while raising the familiar operational and financial questions enterprises must answer before making large-scale cloud storage commitments.
Source: Neowin
Public Preview of Azure Native Dell PowerScale now available