High Performance

High Performance Computing on Azure

HPC workloads impose different challenges that cloud storage solutions struggle to overcome :

  • Access to shared data with high speed.
  • Retrieve information without disruption or delay.
  • Eliminate the movement of datasets between disparate systems.
  • Accelerate the speed at which applications can be deployed and simulations can be run.
  • Set up a high performing environment for massive file number and very write-intensive apps.
  • Protect and recover data in case of hardware or software failure.
  • Deploy resources without waste, and scale seamlessly when needed.

Azure NetApp Files is the lifesaver here thanks to the advanced capabilities it offers to HPC performance-intensive and latency-sensitive workloads.

High performance

  • Data can be accessed in parallel on a single storage volume by different compute nodes.
  • Faster deployment cycles and simulation runs.
  • Bare-metal performance to meet the demands of parallel compute requests.
  • High IOPS/throughput and consistent low latency.

High availability

  • Out-of-the-box high availability of 99.99%.
  • Uninterrupted access to the HPC application stack.


  • No need for proactive capacity planning and no overprovisioning.
  • Meet the scale up and down requirements of your workloads with minimal administrative overhead.
  • Dynamic Provisioning: online scalability of capacity and performance.

Data resiliency

  • Every provisioned volume is protected from hardware, software, or network failures.
  • Space-efficient and time-efficient snapshot-based backup with restore and cloning capabilities.
  • Regional DR providing efficient replication and readable target by default.

End-to-end security

  • Volumes are encrypted at rest using the FIPS 140-2 standard.
  • Data traffic between NFSv4.1 clients and ANF volumes can be encrypted using Kerberos with AES-256 encryption.
  • A unique XTS-AES-256 data encryption key is generated for each volume and never displayed in an unencrypted format.
  • NFS export policy rules to control access to ANF service mount target.
  • Deployment in a delegated subnet within customer’s VNet with no public IP address.

Ease of Management

  • Single management pane for compute, storage and network components of HPC applications.
  • ANF can be provisioned and managed from the Azure portal or through automation scripts like any other Azure service.

Without Azure NetApp Files, you would end up using Parallel File System solutions, whether managed or orchestrated, even for entry level workloads and this is obviously not profitable.

Azure NetApp Files can fit with small to medium size HPC environments, when single-digit or double-digit GB/s throughput is required.

Typical customer deployments of Azure NetApp Files for HPC workloads range from 500 to 5000 computer nodes.

Leave a Reply