Skip to main content
Concept guide · ISO 26262-6 Annex D · 12 chapters

Hypervisors & Mixed-Criticality Systems

When a QM infotainment stack and safety-relevant functions share one SoC, ISO 26262 asks for evidence of freedom from interference. Learn what a hypervisor and the silicon can isolate, what they cannot, and how to verify the claim.

Chapters
12
Chapters
Figures and interactive visuals
15
Figures and interactive visuals
Isolation dimensions
2
Isolation dimensions
The short version

Quick answers

What is a mixed-criticality system in ISO 26262?
A mixed-criticality system runs software with different safety integrity obligations, for example a QM infotainment stack and an ASIL-rated cluster function, on the same computing hardware. ISO 26262 allows this coexistence only if freedom from interference is shown; otherwise all software must be developed to the highest ASIL involved. ISO 26262-6, Annex D gives informative examples of interference grouped by timing and execution, memory, and exchange of information.
How does a hypervisor provide freedom from interference?
It contributes to two dimensions. Spatial isolation keeps each partition inside its own memory and devices, using the second stage of address translation and an I/O memory management unit for DMA-capable devices. Temporal isolation uses scheduling and shared-resource controls so a misbehaving partition cannot starve another of CPU time. Neither is complete on its own: shared caches, buses and memory bandwidth can still couple partitions, and communication between partitions needs receiver-side protection.
Why this course · ISO 26262, Part 6

Why it pays for itself

Isolation you can argue in a safety case

Learn how two-stage address translation, controlled device access and time partitioning become a freedom-from-interference argument an assessor can follow down to the evidence.

Read a hypervisor safety claim critically

Hypervisors offered for safety use come with assumptions of use the integrator must close. Learn which ones matter and where shared resources create coupling the claim does not cover.

Choose by pattern, not by logo

Compare monolithic, microkernel and static-partitioning designs against real deployment archetypes, and decide whether a hypervisor, a safety island or both fit your system.

After the course

What you’ll be able to do

Argue freedom from interference

Build the spatial, temporal and communication argument for software sharing one SoC.

Use the silicon correctly

Know what address translation, device assignment and privilege levels contribute to isolation.

Design partition communication

Exchange data between partitions with receiver-side protection.

Evaluate a hypervisor

Read a safety claim critically and ask the right selection questions.

Verify isolation claims

Plan integration tests and close the assumptions of use behind the claim.

The curriculum · 12 chapters

Chapter by chapter

  1. 01

    Why Virtualize the Car?

    See why E/E architectures consolidate controllers onto shared computers and what mixed criticality really means.

    • From many controllers to shared computers
    • Meaning of mixed criticality
    • One cockpit SoC, three worlds
  2. 02

    Virtualization Fundamentals

    Learn what a hypervisor is, the Type 1 and Type 2 split, paravirtualization, and the isolation spectrum from processes to silicon.

    • Type 1 and Type 2
    • Full versus paravirtualization
    • Isolation spectrum
  3. 03

    What the Silicon Provides

    Understand privilege levels, two-stage address translation and the ways a virtual machine can be given a device.

    • Privilege levels
    • Two-stage address translation
    • Device assignment options
  4. 04

    What ISO 26262 Actually Requires

    Learn why coexistence needs evidence, how the Annex D fault examples apply, and how the partitioning software itself is treated.

    • Coexistence demands evidence
    • Annex D fault examples
    • Interference as dependent failure
  5. 05

    Spatial Isolation

    Treat the memory map as controlled safety evidence and explore which interference paths are blocked and by what.

    • Memory map as evidence
    • Interference-path explorer
    • Beyond RAM
  6. 06

    Temporal Isolation

    Separate the two timing problems, compare how hypervisors hand out CPU time, and map shared-resource interference.

    • Two timing problems
    • Scheduling models
    • Shared-resource interference
  7. 07

    Safe Communication Between Partitions

    See how data crosses a partition boundary and why the receiver, not the channel, has to protect itself.

    • Crossing the boundary
    • Receiver protects itself
    • Patterns and classic mistakes
  8. 08

    The Safety Concept of a Virtualized ECU

    Layer the monitors, design a fault-reaction ladder, secure the boot order and account for FTTI across partitions.

    • Layers of monitoring
    • Fault-reaction ladder
    • FTTI across partitions
  9. 09

    The Hypervisor Landscape

    Compare architectural patterns rather than products, decode a vendor safety claim, and work through selection questions.

    • Patterns, not logos
    • Decoding a safety claim
    • Selection questions
  10. 10

    Deployment Patterns in Real Vehicles

    Study three deployment archetypes, the hypervisor plus safety island combination, and AUTOSAR stacks as guests.

    • Three archetypes
    • Hypervisor plus safety island
    • AUTOSAR as guests
  11. 11

    Verifying the Isolation Claims

    Decide what is being verified, challenge the isolation with integration tests, and close the assumptions of use.

    • Three objects of verification
    • Integration test catalogue
    • Assumptions of use
  12. 12

    Pitfalls, Trade-offs & Decision Guide

    Recognize how virtualized ECUs go wrong and decide whether your system needs a hypervisor at all.

    • Ways virtualized ECUs fail
    • Do you need a hypervisor?
    • Summary in six lines

Read all 12 chapters with Expert.

Start with a free account: the complete Functional Safety Concept guide and 5 practice exams a month, then upgrade when you need this guide.

Expert plan pricing
Diagrams & visuals

Not just text: the visual toolkit

Cockpit SoC with and without partitioning

One SoC hosting several worlds, before and after a partitioning layer.

Two-stage address translation walk

How guest addresses are translated to physical memory.

Interference-path explorer

Access paths between partitions and the mechanisms that block them.

Boot-race timeline

Why isolation must be configured before an untrusted guest starts.

FTTI budget across partitions

Detection, cross-VM signaling, decision and reaction within one interval.

Built for

Who this guide is for

  • Architects consolidating multiple ECUs onto one cockpit or zonal SoC
  • Software engineers whose safety partition now shares silicon with a rich OS
  • Safety engineers writing the freedom-from-interference argument for a virtualized ECU
  • Teams evaluating hypervisor options for an ASIL project

Frequently Asked Questions

Common questions about Hypervisors & Mixed-Criticality Systems

Yes, and digital cockpits are the common example. It works when the safety function runs in its own partition, the partitioning layer is itself developed and verified for the integrity it supports, isolation is configured before any untrusted guest starts, and the remaining shared-resource effects are analyzed as dependent failures. The guide compares this with designs that place the safety function on a separate safety island.
Isolation claims have to become evidence, not datasheet statements. Verification combines integration tests that deliberately challenge the boundary, such as illegal accesses, overruns and shared-resource stress while the safety partition is observed, with a documented closure of every assumption of use in the hypervisor safety manual. Formal proofs can strengthen parts of the argument but do not cover everything. The guide separates the three distinct objects that are being verified.
The guide has 12 chapters: why vehicles consolidate onto shared computers, virtualization fundamentals, what the silicon provides, what ISO 26262 requires, spatial isolation, temporal isolation, communication between partitions, the safety concept for a virtualized ECU, a survey of hypervisor designs, deployment patterns, verifying isolation claims, and a pitfalls and decision guide. It includes 15 figures and interactive visuals. A free account starts you off, and the Pro and Expert plans unlock more of the library.

Start the course today

A free account unlocks one full concept guide, 3 work product templates, 1 guided process, the Markov and RBD simulators, and 5 practice exams per month. The Pro and Expert plans unlock more of the 93-guide library. No credit card required.

Continue with

Where this topic leads next.

Browse all 93 guides
  • System levelMulti-Core Safety ArchitecturesHow automotive silicon catches its own random hardware faults: lockstep and split-lock cores, safety islands on big SoCs, and hardware and software self-tests. Then how you turn all of it into an ASIL claim with real coverage numbers.Expert plan pricing12 chapters
  • System levelPower & Clock SupervisionThe two assumptions every other safety mechanism silently depends on: valid supply rails and a trustworthy time base. Power trees, safety PMICs, voltage monitors, sequencing, reset escalation, watchdogs, clock monitors, the independence argument, timing and energy budgets, latent-fault self-test, and the FMEDA rows they all feed.Expert plan pricing14 chapters
  • HardwareHardware Engineering in ISO 26262Learn the complete hardware development lifecycle per ISO 26262-5 from requirements through design and verification. Covers HSI specification, SPFM/LFM/PMHF achievement, dependent failure analysis, and ASIC/FPGA/PCB considerations.Expert plan pricing9 chapters