Skip to content

Security: Vector-Pi/defence-middleware

Security

SECURITY.md

Security

Scope

This is a simulation-based research proof of concept. It is not intended for operational deployment in any form. The security considerations below are documented for completeness and to support readers of the accompanying paper.

Known security limitations

No authentication. MIL-STD-1553B embeds an implicit trust model in which physical access to the bus cable constitutes full authorisation. STANAG 4586 provides an LOI-based authorisation handshake (Messages 1, 20, 21) but no cryptographic proof of identity. Neither is implemented in this proof of concept. Gap G-013 in the paper classifies this as UNTRANSLATABLE.

Network-accessible command path. The middleware introduces a UDP endpoint on port 5400 that accepts STANAG messages and translates them into 1553 bus transactions. In a production system this endpoint must be protected by mutual TLS or IPsec before any operational deployment. This is not implemented here.

Replay protection. The simulation enforces monotonic instance identifiers per STANAG message type (Gap G-015), which provides application-layer replay detection under normal conditions. It does not protect against an adversary who can observe and replay traffic within a single session.

Loopback only. The proof of concept binds to 127.0.0.1. It does not listen on any external interface. There is no exposure beyond the local machine when running in its default configuration.

Reporting

This is a research repository. If you identify a security issue relevant to the architecture described in the paper, contact:

om@omarora.in

Please do not open a public GitHub issue for security concerns.

There aren’t any published security advisories