Organizations are increasingly looking to migrate on-premises NFS file storage to the cloud to reduce infrastructure costs, simplify operations, and unlock new opportunities for innovation. However, Linux workloads have diverse storage requirements. Some prioritize cost efficiency, while others demand significantly greater scale and performance, making it challenging to find a cloud file storage solution that meets both needs.
At Microsoft Ignite, we're expanding the capabilities of our existing Azure Files HDD and SSD tiers to address both challenges. NFS support on Azure Files HDD provides a more cost-effective path for migrating traditional Linux workloads, helping customers reduce total cost of ownership (TCO). At the same time, scale performance on Azure Files SSD raises the limits of a single NFS share to 1 PiB of capacity, 1 million IOPS, and 20 GiB/s of throughput, enabling larger, more demanding workloads.
Together, these enhancements expand the range of on-premises workloads customers can migrate to Azure Files, from traditional enterprise applications to high-performance computing (HPC), scientific computing, and AI/ML. Beyond migration, Azure Files also provides a foundation for unlocking the value of enterprise data through AI-powered experiences and scalable shared storage for AI applications and agents.
NFS SSD scale performance: Unlock larger, more demanding workloads
Scale and performance have historically been barriers to migrating large, data-intensive workloads from on-premises environments to the cloud. Applications such as HPC, genomics, engineering simulations, and AI/ML often process massive datasets across hundreds or thousands of compute processes simultaneously.
Traditional file-share limits can require organizations to partition datasets across multiple file systems, increasing operational complexity and limiting application scalability.
With scale performance on Azure Files SSD, we're significantly increasing the capabilities of a single NFS share:
- Up to 1 PiB of capacity to accommodate massive datasets without splitting data across multiple shares.
- Up to 1 million IOPS to support highly parallel, I/O-intensive workloads.
- Up to 20 GiB/s of throughput to accelerate data-intensive processing across distributed compute environments.
To get started, complete the scale performance SSD signup survey.
High-performance computing: Support larger datasets and parallel processing
HPC workloads such as genomics, seismic processing, and scientific computing often involve massive datasets accessed concurrently by hundreds or thousands of compute processes.
For example, geoscientists rely on large seismic datasets to understand the Earth's subsurface. Before these datasets can be interpreted, they often undergo intensive processing across hundreds of virtual machines reading and writing data in parallel, making storage throughput critical.
With greater capacity, IOPS, and throughput on a single share, Azure Files SSD can support larger datasets and greater processing parallelism. This can help reduce storage bottlenecks in seismic processing, enabling geoscientists to move quickly from raw data to 3D interpretation, where they analyze geological structures, seismic slices, and well data to identify potential carbon storage formations.
By enabling larger datasets and greater parallelism without the complexity of managing multiple file shares, Azure Files SSD helps organizations migrate and run a broader range of HPC and data-intensive workloads in Azure.
NFS HDD: Reduce TCO and simplify on-premises file migration
Not every Linux workload requires the performance of SSD storage. Many traditional enterprise applications rely on NFS file shares for large datasets, application data, and general-purpose storage, where cost efficiency is often more important than high performance.
With the introduction of NFS support on Azure Files HDD, we're providing customers with a lower-cost storage option for migrating these workloads to Azure.
Organizations can retain familiar NFS access for existing Linux applications while avoiding the additional expense of SSD performance they don't need.
By combining cost-effective HDD storage with a fully managed file service, Azure Files helps customers reduce infrastructure costs, simplify storage operations, and optimize TCO when migrating existing NFS workloads to the cloud.
To get started, complete the NFS HDD signup survey.
AI-ready: Unlock the value of enterprise file data
Organizations have accumulated years of valuable data across on-premises file servers and NAS environments, including business documents, research datasets, engineering files, and application data. Bringing this data to Azure provides a foundation for building AI-powered applications and unlock that benefits from that data.
Azure Files supports this journey in two ways.
Connect enterprise data to AI-powered experiences: Through the Azure File Share Microsoft 365 and Logic Apps connector, organizations can integrate file-based data into AI solutions built with Microsoft Foundry and Microsoft 365 Copilot. These integrations enable enterprises to unlock the value of their existing data by making it discoverable, searchable, and accessible to AI-powered applications and agents.
Scaling shared NFS storage for AI applications and agents: Applications running on Azure Kubernetes Service (AKS) already use Azure Files NFS as shared storage. AI applications can store and process shared datasets, while agents can use persistent file storage to retain their working data between runs. With SSD scale performance, a single SSD-backed NFS share can hold more data and handle more reads and writes in parallel as these workloads grow.
Get started with Azure Files NFS
NFS shares with scale performance on SSD and NFS HDD are available in both the classic and new file share experiences.
Both capabilities are supported exclusively with the NFS protocol on the Provisioned v2 billing model. Scale performance is available on SSD, while NFS HDD extends NFS support to the HDD tier.
To get started, complete the corresponding surveys to request access, and the Azure Files team will contact you with next steps.
For questions, contact azurefiles@microsoft.com.