flexFS Enterprise - Elastic POSIX Network Filesystem

flexFS is a high-performance network filesystem purpose-built for I/O-intensive applications where throughput, cost, and scale all matter. From GPU training clusters to genomics pipelines, teams use flexFS to eliminate storage bottlenecks without changing their code. Applications read and write files using standard filesystem calls while flexFS handles block storage, caching, compression, and encryption transparently.

All file system services, metadata, and data reside entirely within your own account, and on infrastructure you exclusively provision and control. Only billing (required) and support (optional) telemetry ever leaves your environment. You decide if, and when, you want to take newer releases.

Replace provisioned cloud filesystems such as EFS and FSx with metered object storage. Pay only for what you store, not for idle capacity. Organizations have reported projected savings of millions of dollars over multi-year periods and storage cost reductions of over 50% on individual projects.

By default, mount clients access storage directly, so aggregate read throughput scales linearly with every client you add. There are no data-path bottlenecks, and flexFS supports exabyte-scale capacity. Add more compute nodes and your storage performance grows with them. When configured, flexFS proxy groups can scale horizontally to add high-performance writeback caching and acceleration to flexFS clusters and mitigate object storage latency.

flexFS delivers full POSIX compliance including ACLs, extended attributes, file locking, hard links, symlinks, and special files. Your applications, scripts, and tools work unchanged.

When enabled, AES-256-GCM encryption with Argon2id key derivation keeps your data secure end-to-end. Keys are derived on the client and never leave your hosts. The metadata server, proxies, and storage backends never see plaintext data.

• Proxy Groups: CDN-like caching layer with automatic lowest-latency proxy selection and writeback caching for hybrid cloud deployments
• Time-Travel Mounting: Mount any volume at any point in time using read-only access for auditing, recovery, and compliance - no separate snapshot infrastructure required
• Zero-Downtime Updates: Seamless session handoff lets you update mount clients without unmounting or interrupting running workloads
• Unlimited Volumes: Enterprise Edition removes the single-volume and 5 TiB limits of the Community Edition
• Dynamic Kubernetes Provisioning: CSI volume driver with Helm chart deployment supporting both static and dynamic provisioning through standard PersistentVolumeClaims

flexFS includes a CSI volume driver with Helm chart for easy deployment. Static and dynamic provisioning let pods mount object storage directly through standard PersistentVolumeClaims, making it straightforward to integrate into existing container orchestration workflows on Amazon EKS.

Organizations across pharmaceuticals, government research, and other industries rely on flexFS for their most demanding workloads. Teams at top pharma companies and NASA contractors have reported significant cost savings and performance improvements after adopting flexFS.