Skip to content

AFDX / ARINC 664

Conceptual virtual-link observation surface for ARINC 664 / AFDX teaching.

Status Examples Runtime example Source posture
draft-core-lab synthetic/passive available standards metadata plus public/tooling context

Identity

ARINC 664 Part 7 / AFDX is represented here only as conceptual virtual-link observation evidence: virtual link, redundant network side, sequence, observed time, length, payload fingerprint, and apparent cadence.

What This Protocol Teaches

  • AFDX constrains Ethernet with avionics timing and redundancy rules, but this page does not implement Ethernet inspection.
  • Virtual links move authority from free-form packets to configured communication channels.
  • Observers reason about apparent BAG/cadence, sequence, redundant A/B evidence, and missing counterparts.

Operational Context

ARINC 664 Part 7 / AFDX appears in commercial-aviation, avionics contexts and is modeled in the atlas at the data-link, network layer(s). The binder treats it as a passive inspection surface: what virtual-link evidence appears, what timing matters, and what outside authority is required before assigning payload meaning.

This page intentionally does not parse raw Ethernet, UDP, IP, or PCAP data. It has no switch simulation, routing, end-system scheduling, policing, or conformance behavior.

Draft status

draft-core-lab: this page is in Core Lab scope but is not yet promoted to final copy. It has a source-grounded artifact and review structure; it still needs the same page-specific depth as CAN, DBC, MIL-STD-1553, UART, ARINC 429, and Modbus RTU before final-copy status.

A Concrete Artifact

# synthetic AFDX-like observation
VL 0x0123  seq=0x5a  network=A  len=64  t=12.000 ms
VL 0x0123  seq=0x5a  network=B  len=64  t=12.001 ms

Synthetic

Synthetic offline sample for explanation; not a real operational trace or live-system instruction.

Worked Decode

  1. Both observations claim the same virtual link and sequence on redundant networks.
  2. The near-identical timing suggests duplicate delivery over A/B paths, not two independent application updates.
  3. The virtual link configuration and BAG assumption determine how to assess apparent cadence in this synthetic observation.
  4. Payload meaning remains application-profile authority.

Field Layout / Anatomy

Element Shape Inspection meaning
Virtual link configured identifier Communication channel with bandwidth and destination set.
BAG configured interval Minimum spacing authority for a VL.
Sequence number integrity/ordering evidence Helps detect loss or duplicate behavior.
A/B networks redundancy Independent paths for availability.
Payload fingerprint observer summary Supports same/different payload evidence without decoding application meaning.

Visual Model

flowchart LR lru1["End system"] --> swA["Network A switch"] lru1 --> swB["Network B switch"] swA --> lru2["Receiver"] swB --> lru2 cfg["Virtual link config / apparent BAG"] --> swA cfg --> swB

Timing And Authority

Authority comes from configuration: virtual-link allocation, BAG, and redundancy management. Semantic authority comes from the avionics application carried over the VL. A passive observer can compare apparent cadence, sequence continuity, A/B skew, and payload fingerprints, but cannot prove receiver acceptance, aircraft state, or application semantics.

Semantic authority

Bytes rarely explain themselves. Name the source that defines meaning before naming engineering values: standard metadata, label table, DBC, PGN/SPN definition, object dictionary, register map, DID catalog, LDF, ARXML, channel metadata, or vendor profile.

Failure And Ambiguity

  • Treating redundant A/B duplicates as separate state updates doubles the apparent rate.
  • Seeing only A or only B is a missing counterpart in the observation, not proof of path failure.
  • Duplicate sequence evidence does not prove consuming application behavior.
  • Apparent BAG/cadence evidence is a teaching assumption here, not conformance checking.
  • Ethernet/IP familiarity can mislead observers into assuming open network behavior.
  • Application data remains opaque without interface-control authority.

Observer Lesson

A passive observer can infer conceptual virtual-link cadence, possible missing observations, duplicate sequence evidence, skew between redundant networks, and missing A/B counterparts. It cannot infer application semantics, receiver acceptance, physical state, or safety effect.

Python Model

schoolbus exposes a conceptual virtual-link observation model:

"""Runnable AFDX conceptual virtual-link observation example."""

from datetime import UTC, datetime, timedelta

from schoolbus.protocols.afdx import (
    AfdxVirtualLinkObservation,
    assess_bag_cadence,
    assess_sequence_evidence,
    compare_redundant_observations,
)

base = datetime(2026, 5, 21, 12, 0, tzinfo=UTC)

network_a = AfdxVirtualLinkObservation(
    virtual_link_id=0x123,
    network="A",
    sequence=0x5A,
    observed_at=base,
    length=64,
    payload_fingerprint="sha256:synthetic-vl-0123",
    source="samples/avionics-workbook",
)
network_b = AfdxVirtualLinkObservation(
    virtual_link_id=0x123,
    network="B",
    sequence=0x5A,
    observed_at=base + timedelta(milliseconds=1),
    length=64,
    payload_fingerprint="sha256:synthetic-vl-0123",
    source="samples/avionics-workbook",
)
network_a_next = AfdxVirtualLinkObservation(
    virtual_link_id=0x123,
    network="A",
    sequence=0x5B,
    observed_at=base + timedelta(milliseconds=4),
    length=64,
    payload_fingerprint="sha256:synthetic-vl-0123-next",
    source="samples/avionics-workbook",
)

print(network_a.explain())
print(
    compare_redundant_observations(
        network_a,
        network_b,
        max_skew=timedelta(milliseconds=2),
    ).explain()
)
successive_network_a = (network_a, network_a_next)
print(assess_sequence_evidence(successive_network_a).explain())
print(
    assess_bag_cadence(
        successive_network_a,
        expected_bag=timedelta(milliseconds=4),
    ).explain()
)

The model exposes virtual-link evidence, apparent cadence, sequence evidence, and A/B comparison. It deliberately has no raw Ethernet, no UDP, no IP, no PCAP parser, no switch simulation, and no conformance checker.

The avionics workbook includes passive synthetic conceptual virtual-link records under samples/avionics-workbook/.

Local Teaching Notes

Simplification

The observation is synthetic and omits exact Ethernet/IP/UDP encapsulation details.

Source Confidence

Source confidence is high for scope and terminology, with semantic claims limited to public metadata, public source notes, and synthetic teaching artifacts.

Source Notes

Field Value
Governance tier Tier 1 Core Lab
Canonical source status yes
Public explainer status yes
Open-source tool status yes
Sample-data status none listed; use synthetic teaching artifacts
Confidence high
Citation specificity document-metadata-level
Canonical source(s) ARINC 664 Part 7 Avionics Full-Duplex Switched Ethernet Network (SAE ITC / ARINC; ARINC 664 Part 7; paywalled; metadata-only)
Public explainer/tooling source(s) AFDX / ARINC 664 Part 7 tutorial (AIM Online; public-web; link-only)
Open-source tool references Wireshark: public dissector/tooling context; not used here as a parser surface

References

Public Sources