AWS has introduced KNFSD File Cache, an open-source NFS caching proxy designed for hybrid and burst-compute workloads that cannot immediately move their primary storage.
The design mounts an existing NFS export and re-serves it to cloud clients. Frequently used data is held first in the Linux block cache and then in a persistent local-NVMe tier. AWS provides infrastructure guidance for autoscaling, load distribution, metrics and hardened images on x86 or Graviton systems.
Why it matters
Media pipelines often scale compute more easily than storage. Rendering, transcode and AI jobs may burst into the cloud, but the assets remain on a facility filer. Copying an entire project can be slower and more expensive than the job itself.
A transparent cache can reduce that friction because clients keep using NFS semantics. It does not remove planning: write-back behaviour, cache consistency, source availability, working-set size, egress cost and security still need production testing. Workloads with heavy random writes may not see the same benefits as read-dominant rendering or processing.
