FileDNA Server
Enforcement before a file reaches anyone’s desktop.
FileDNA Server runs the CADR engine as a persistent, always-on daemon — inspecting files in transit or at rest across mail flow, file shares, and upload pipelines, before a single endpoint ever sees them.
How It Works
Server keeps the engine warm and serves requests over a local wire protocol, so enforcement doesn’t cost a fresh process spin-up per file.
1
Load once
The daemon loads the FileDNA engine a single time at startup and keeps it warm, avoiding the process-spawn cost of invoking it fresh on every file.
2
Listen
Clients — a mail gateway, an upload handler, a file-share watcher — connect over a local or network socket and submit a file path and a desired mode.
3
Process on a dedicated worker
Each request is served by an isolated worker process, so a malformed or hostile file can’t destabilize the daemon itself.
4
Respond and, optionally, replace
The daemon returns a structured verdict. In reconstruct mode, the cleaned file is written to a temporary path, fsync’d, and atomically renamed over the original — nothing downstream ever reads a half-written file mid-swap.
Key Capabilities
Designed for the chokepoints where per-endpoint enforcement doesn’t scale.
Origin-aware policy, not one-size-fits-all
Route by where a file came from — a trusted internal share can stay scan-only, while a downloads folder or an inbound mail queue defaults to reconstruct — through a configurable mode-selection policy.
Built for throughput
A warm worker pool means the engine’s load cost is paid once, not per file, which is what makes server-side, inline enforcement realistic at mail-gateway or file-share volume.
Fails the way you choose
Fail-closed by default for deployments where an unavailable engine should block rather than silently pass a file through; fail-open is available as an explicit, documented opt-in for availability-first deployments.
Atomic by construction
Every reconstruction is written to a temp file and swapped in with a single atomic rename — there’s no window where a partially-written file is readable.
Who It’s For
Security and IT teams enforcing file hygiene at a chokepoint — mail gateway, file-share ingestion, upload pipeline — rather than at every individual endpoint.
Organizations that need centralized policy and centralized logging for file-borne threat handling, not a per-machine deployment.
Teams already running FileDNA Desktop who want the same engine enforced server-side as a second layer, not a replacement.