Contents
SpecificationAUTOSAR

AUTOSAR E2E protection

End-to-end communication protection

AUTOSAR end-to-end protection puts a counter, a data identifier, and a CRC in the payload so the receiver can tell when safety-related data was corrupted, repeated, lost, reordered, or confused with another signal.

What it is

End-to-end (E2E) communication protection is the AUTOSAR mechanism that lets a receiver decide whether safety-related data still matches what the sender application produced. The check travels in the payload from the sending software component to the receiving one, so a fault introduced in the RTE, in COM, in a driver, or on the network stays visible to the software that uses the value.

The sender adds a rolling counter and a CRC. The CRC covers the application data and a data identifier agreed for that signal. The receiver recomputes the CRC, confirms the identifier, and checks that the counter advanced by an allowed step. Profiles fix the header layout and which of these checks apply. Classic Platform supplies the profiles as a BSW library and as an RTE transformer. Adaptive Platform communication management attaches the check result to the received sample. The library returns a status, and the application decides whether the value may be used.

Who it applies to

E2E applies to AUTOSAR Classic and Adaptive ECUs that exchange safety-related signals. A project makes it mandatory by selecting an AUTOSAR baseline that includes these documents.

Safety engineers choose which signals are protected and which profile the safety concept requires. BSW integrators set the profile, the data identifiers, and the call site: the application, a COM wrapper, or the RTE transformer. Application developers protect a signal once and, on reception, act on the check result. On a high-performance computer that result is often consumed on the safety-island MCU, by platform services and by applications that use data from another core, another ECU, or a vehicle bus only after the check passes.

What it covers

The specification defines the protect step, the check step, the receiver state machine, and the configuration of each profile.

Fault What the receiver uses
Payload corruption CRC over the data and the data identifier
Wrong signal or wrong layout Data identifier included in the CRC
Repetition, loss, insertion, or wrong order Rolling counter and an allowed delta
No fresh sample Receiver state machine when no new sample arrives

Profiles differ in header size and in which of these faults they cover. The project selects a profile that fits the PDU and the safety requirement. Frame layout on CAN, CAN FD, FlexRay, and Ethernet remains in the network specifications.

How it fits with other documents

ISO 26262 expects the safety concept to address faults in the exchange of safety-related data, including corruption, repetition, loss, delay, insertion, masquerading, and incorrect sequence. E2E is the AUTOSAR measure safety analyses usually name for that exchange.

Classic Platform places the call in the application, in a COM protection wrapper, or in an RTE E2E transformer that returns the status with the data element. Adaptive communication management uses the same profile model for events and fields. The network carries the protected bytes and leaves interpretation to the E2E check.

Secure Onboard Communication supplies freshness and an authenticator for the same signal when the cybersecurity concept requires authenticity. The two concepts set the header order. Network management, diagnostics, and software update can withhold a fresh sample; the E2E state machine then reports the data as unchecked.

In practice

Give every protected signal or signal group a data identifier that no other signal at that receiver shares. Both sides use the same profile, the same identifier, and the same counter rules. A duplicated identifier, or counters that start out of step, fails on every reception until the configuration matches.

On the safety island, applications and platform services read the E2E status before a value enters control or a safety output. A repeated counter is a repeated sample even when the CRC matches. The safety concept sets the reaction for a missing sample and for a failed CRC: hold the last value, enter a safe state, or raise a fault.

Integration checks that protection runs in exactly one place, either the application, COM, or the transformer, and that every signal on the safety list maps to a profile. A pass shows a matching counter and CRC. A cryptographic key check stays with SecOC.

What it means for your discipline

  • Functional safetyCore

    Safety engineers list the signals whose safety concept depends on detecting corruption, repetition, loss, insertion, wrong sequence, and masquerading, and they assign the profile that covers those faults. They also define the application reaction for each unsuccessful status, including a missing sample.

  • Base softwareCore

    BSW integrators configure the profile, the data identifier, and the allowed counter step, and they choose whether the application, a COM wrapper, or the RTE transformer executes protect and check. Mismatched identifiers or two layers protecting the same signal are integration faults they own.

  • HPC & application softwareCore

    Application developers protect each outgoing safety signal once and branch on the E2E status before a received value enters control or a safety output. On a safety-island MCU this includes platform services that qualify data arriving from another core or from the vehicle network.

  • E/E architecturePartial

    Architects decide which safety-related signals cross an ECU or partition boundary and which end performs the check, including the safety island on a high-performance computer. They place E2E on that path and leave cryptographic authentication to the cybersecurity concept when forgery is in scope.

  • Networks & communicationPartial

    Network and gateway engineers keep the protected header intact while routing and size each PDU so the selected profile fits the frame. Dropped or delayed frames still surface at the receiver as a failed or missing E2E check.

  • CybersecurityPartial

    Cybersecurity engineers treat the E2E CRC as a safety checksum that a sender who knows the profile can recompute. When the threat analysis requires authenticity, they add Secure Onboard Communication and agree the header order with the safety team.

No direct stake: Hardware & silicon · Diagnostics & lifecycle · Software update & OTA