AzureFeeds

Disclaimer: The following is a post authored by our partner Qumulo. Qumulo has been a valued partner in the Azure Storage ecosystem for many years and we are happy to share details on their unique approach to solving challenges of scalable filesystems!

Whether you’re training massive AI models, running HPC simulations in life sciences, or managing unstructured media archives at scale, performance is everything. Qumulo and Microsoft Azure deliver the cloud-native file system built to handle the most data-intensive workloads, with the speed, scalability, and simplicity today’s innovators demand.

But supporting modern workloads at scale is only part of the equation. Qumulo and Microsoft have resolved one of the most entrenched and difficult challenges in modernizing the enterprise data estate: empowering file data with high performance across a global workforce without impacting the economics of unstructured data storage.

According to Gartner, global end-user spending on public cloud services is set to surpass $1 trillion by 2027. That staggering figure reflects more than just a shift in IT budgets—it signals a high-stakes race for relevance.

CIOs, CTOs, and other tech-savvy execs are under relentless pressure to deliver the capabilities that keep businesses profitable and competitive. Whether they’re ready or not, the mandate is clear: modernize fast enough to keep up with disruptors, many of whom are using AI and lean teams to move at lightning speed. To put it simply, grow margins without getting outpaced by a two-person startup using AI in a garage. That’s the challenge leaders face every day.

Established enterprises must contend with the duality of maintaining successful existing operations and the potential disruption to those operations by a more agile business model that offers insight into the next wave of customer desires and needs. Nevertheless, established enterprises have a winning move – unleash the latent productivity increases and decision-making power hidden within years, if not decades, worth of data. Thoughtful CIOs, CTOs, and CXOs have elected to move slowly in these areas due to the tyranny of quarterly results and the risk of short-term costs reflecting poorly on the present at the expense of the future. In this sense, adopting innovative technologies forced organizations to choose between self-disruption with long-term benefits or non-disruptive technologies with long-term disruption risk. When it comes to network-attached storage, CXOs were forced to accept non-disruptive technologies because the risk was too high.  

This trade-off is no longer required. Microsoft and Qumulo have addressed this challenge in the realm of unstructured file data technologies by delivering a cloud-native architecture that combines proven Azure primitives with Qumulo’s suite of file storage solutions. Now, those patient CXOs, waiting to adopt hardened technologies, can shift their file data paradigm into Azure while improving business value, data portability, and reducing the financial burden on their business units. 

Today, organizations that range from 50,000+ employees with global offices, to organizations with a few dozen employees with unstructured data-centric operations have discovered the incredible performance increases, data availability, accessibility, and economic savings realized when file data moves into Azure using one of two Qumulo solutions:

Option 1 — Azure Native Qumulo (ANQ) is a fully managed file service that delivers truly elastic capacity, throughput, and IOPS, along with all the enterprise features of your on-premises NAS and a TCO to match.

Option 2 — Cloud Native Qumulo (CNQ) on Microsoft Azure is a self-hosted file data service that offers the performance and scale your most demanding workloads require, at a comparable total cost of ownership to on-premises storage.

Both CNQ on Microsoft Azure and ANQ offer the flexibility and capacity of object storage while remaining fully compatible with file-based workflows. As data platforms purpose-built for the cloud, CNQ and ANQ provide three key characteristics:

1. Elasticity — Performance and capacity can scale independently, both up and down, dynamically.


2. Boundless Scale — Virtually no limitations on file system size or file count, with full multi-protocol support.


3. Utility-Based Pricing — Like Microsoft Azure, Qumulo operates on a pay-as-you-go model, charging only for resources used without requiring pre-provisioned capacity or performance.

The collaboration between Qumulo’s cloud-native file solutions and the Microsoft Azure ecosystem enables seamless migration of a wide range of workflows, from large-scale archives to high-performance computing (HPC) applications, from on-premises environments to the cloud. For example, a healthcare organization running a fully cloud-hosted Picture Archiving and Communication System (PACS) alongside a Vendor Neutral Archive (VNA) can leverage Cloud Native Qumulo (CNQ) to manage medical imaging data in Azure. CNQ offers a HIPAA-compliant, highly durable, and cost-efficient platform for storing both active and infrequently accessed diagnostic images, enabling secure access while optimizing storage costs.

With Azure’s robust cloud infrastructure, organizations can design a cloud file solution that scales to meet virtually any size or performance requirement, while unlocking new possibilities in cloud-based AI and HPC workloads. Further, using the Qumulo Cloud Data Fabric, the enterprise is able to connect geographically separated data sources within one unified, strictly consistent (POSIX-compliant), secure, and high-performance file system.

As organizational needs evolve — whether new workloads are added or existing workloads expand — Cloud Native Qumulo or Azure Native Qumulo can easily scale to meet performance demands while maintaining the predictable economics that meet existing or shrinking budgets.

About Azure Native Qumulo and Cloud Native Qumulo on Azure

Azure Native Qumulo (ANQ) and Cloud Native Qumulo (CNQ) enable organizations to leverage a fully customizable, multi-protocol solution that dynamically scales to meet workload performance requirements. Engineered specifically for the cloud, ANQ is designed for simplicity of operation and automatic scalability as a fully managed service. CNQ offers the same great technology, directly leveraging cloud-native resources like Azure Virtual Machines (VMs), Azure Networking, and Azure Blob Storage to provide a scalable platform that adapts to the evolving needs of today’s workloads – but deploys entirely in the enterprise tenant, allows for direct control over the underlying infrastructure, and requires a little bit higher level of internal expertise to operate.

Azure Native Qumulo and Cloud Native Qumulo on Azure also deliver a fully dynamic file storage platform that is natively integrated with the Microsoft Azure backend. Here’s what sets ANQ and CNQ apart:

• Elastic Scalability — Each ANQ and CNQ instance on Azure Blob Storage can automatically scale to exabyte-level storage within a single namespace by simply adding data. On Microsoft Azure, performance adjustments are straightforward: just add or remove compute instances to instantly boost throughput or IOPS, all without disruption and within minutes. Plus, you pay only for the capacity and compute resources you use.


• Deployed in Minutes — ANQ deploys from the Azure Portal, CLI, or PowerShell, just like a native service. CNQ runs in your own Azure virtual network and can be deployed via Terraform. You can select the compute type that best matches your workload’s performance requirements and build a complete file data platform on Azure in under six minutes for a three-node cluster.


• Automatic TCO Management — can be facilitated through services like Komprise Intelligent Tiering for Azure and Azure Blob Storage access tiers. It optimizes storage costs and manages data lifecycle. By analyzing data access patterns, these systems move files or objects to appropriate tiers, reducing costs for infrequently accessed data. Additionally, all data written to CNQ is compressed to ensure maximum cost efficiency.

ANQ automatically adapts to your workload requirements, and CNQ’s fully customizable architecture can be configured to meet the specific throughput and IOPS requirements of virtually any file or object-based workload. You can purchase either ANQ or CNQ through a pay-as-you-go model, eliminating the need to pre-provision cloud file services. Simply pay for what you use. ANQ and CNQ deliver comparable performance and services to on-premises file storage at a similar TCO.

Qumulo’s cloud-native architecture redefines cloud storage by decoupling capacity from performance, allowing both to be adjusted independently and on demand. This provides the flexibility to modify components such as compute instance type, compute instance count, and cache disk capacity — enabling rapid, non-disruptive performance adjustments. This architecture, which includes the innovative Predictive Cache, delivers exceptional elasticity and virtually unlimited capacity. It ensures that businesses can efficiently manage and scale their data storage as their needs evolve, without compromising performance or reliability.

ANQ and CNQ retain all the core Qumulo functionalities — including real-time analytics, robust data protection, security, and global collaboration.

Example architecture

In the example architecture, we see a solution that uses Komprise to migrate file data from third-party NAS systems to ANQ. Komprise provides platform-agnostic file migration services at massive scale in heterogeneous NAS environments. This solution facilitates the seamless migration of file data between mixed storage platforms, providing high-performance data movement, ensuring data integrity, and empowering you to successfully complete data migration projects from your legacy NAS to an ANQ instance.

Figure: Azure Native Qumulo’s exabyte-scale file data platform and Komprise

Beyond inherent scalability and dynamic elasticity, ANQ and CNQ support enterprise-class data management features such as snapshots, replication, and quotas. ANQ and CNQ also offer multi-protocol support — NFS, SMB, FTP, and FTP-S — for file sharing and storage access. Additionally, Azure supports a wide range of protocols for various services. For authentication and authorization, it commonly uses OAuth 2.0, OpenID Connect, and SAML. For IoT, MQTT, AMQP, and HTTPS are supported for device communication. By enabling shared access to the same data via all protocols, ANQ and CNQ support collaborative and mixed-use workloads, eliminating the need to import file data into object storage. Qumulo consistently delivers low time-to-first-byte latencies of 1–2ms, offering a combined file and object platform for even the most performance-intensive AI and HPC workloads.

ANQ and CNQ can run in all Azure regions (although ANQ operates best in regions with three availability zones), allowing your on-premises data centers to take advantage of Azure’s scalability, reliability, and durability. ANQ and CNQ can also be dynamically reconfigured without taking services offline, so you can adjust performance — temporarily or permanently — as workloads change. An ANQ or CNQ instance deployed initially as a disaster recovery or archive target can be converted into a high-performance data platform in seconds, without redeploying the service or migrating hosted data.

If you already use Qumulo storage on-premises or in other cloud platforms, Qumulo’s Cloud Data Fabric enables seamless data movement between on-premises, edge, and Azure-based deployments. Connect portals between locations to build a Global Namespace and instantly extend your on-premises data to Azure’s portfolio of cloud-native applications, such as Microsoft Copilot, AI Studio, Microsoft Fabric, and high-performance compute and GPU services for burst rendering or various HPC engines. Cloud Data Fabric moves files through a large-scale data pipeline instantly and seamlessly.

Use Qumulo’s continuous replication engine to enable disaster recovery scenarios, or combine replication with Qumulo’s cryptographically locked snapshot feature to protect older versions of critical data from loss or ransomware. ANQ and CNQ leverage Azure Blob’s 11-nines durability to achieve a highly available file system and utilizes multiple availability zones for even greater availability — without the added costs typically associated with replication in other file systems.

Conclusion

The future of enterprise storage isn’t just in the cloud — it’s in smart, cloud-native infrastructure that scales with your business, not against it. Azure Native Qumulo (ANQ) and Cloud Native Qumulo (CNQ) on Microsoft Azure aren’t just upgrades to legacy storage — they’re a reimagining of what file systems can do in a cloud-first world. Whether you’re running AI workloads, scaling HPC environments, or simply looking to escape the limitations of aging on-prem NAS, ANQ and CNQ give you the power to do it without compromise. With elastic performance, utility-based pricing, and native integration with Azure services, Qumulo doesn’t just support modernization — it accelerates it.

To help you unlock these benefits, the Qumulo team is offering a free architectural assessment tailored to your environment and workloads. If you’re ready to lead, not lag, and want to explore how ANQ and CNQ can transform your enterprise storage, reach out today by emailing Azure@qumulo.com. Let’s build the future of your data infrastructure together.