VeriLog
An edge-native evidence engine that turns device telemetry into tamper-evident logs with signed checkpoints. The project focuses on retaining verifiable operational history close to the device that produced it.
Operational history anchored at the edge
VeriLog records device events in an append-only structure and periodically signs checkpoints. The aim is to make later alteration detectable without requiring the edge device to stream every raw event to a central service.
The design is intended for systems where connectivity is intermittent, devices are resource constrained, or central custody would weaken the provenance story.
- 01Append-only event history
- 02Merkle-based integrity structure
- 03Signed checkpoints for later verification
- 04Edge-first operation with delayed export
What evidence survives an untrusted path?
The central question is how much assurance a compact edge log can provide after telemetry has crossed devices, networks, and storage systems that may not all be trusted.
VeriLog explores the boundary between ordinary logging and evidence: ordering, checkpoint cadence, inclusion verification, export, and the information a verifier still needs.
Evidence boundary
VeriLog is an early prototype. Its current claims concern the project architecture and implemented evidence direction, not independent validation across production device fleets.