Case Studies

Automotive OEM ISO 21434 Compliance

An anonymized look at how a major automotive OEM used Safeguard.sh to operationalize ISO/SAE 21434 software supply chain requirements across vehicle platforms.

Shadab Khan
Security Engineer
7 min read

A major automotive OEM shipping passenger vehicles in North America, Europe, and Asia had, by early 2026, completed most of the organizational work required for ISO/SAE 21434 compliance. A Cybersecurity Management System (CSMS) was in place. TARA (Threat Analysis and Risk Assessment) methodology was documented. Cybersecurity engineering practices were integrated into the vehicle development lifecycle. What was missing — and what increasingly appeared in audits and homologation reviews — was a credible operational answer to the software supply chain portions of the standard, particularly those tied to component-level vulnerability awareness across the hundreds of ECUs in a modern vehicle. This is an illustrative account of how Safeguard.sh helped the OEM operationalize the software side of ISO 21434 across vehicle platforms.

Why Is ISO 21434 Software Supply Chain Work So Complex for OEMs?

It is complex because a modern vehicle is not a single software system. It is a federation of dozens to hundreds of software systems, most of which are built by tier-one and tier-two suppliers, each with their own development practices, component sourcing, and security maturity. ISO 21434 expects the OEM to manage cybersecurity risk across this entire distributed system, including the software components inside supplier-delivered ECUs that the OEM did not author and does not always have source code for.

The OEM's Head of Vehicle Cybersecurity described the challenge pragmatically: "We have 120 suppliers on our primary passenger platform. Some of them provide full source code. Some provide nothing but binaries. We need a single operational view of what is in our vehicles and how it is aging." The team had been building this view manually, with inconsistent results, across three vehicle platforms.

Safeguard.sh was engaged to serve as the platform for ingesting, normalizing, and continuously analyzing software composition information from across the supplier ecosystem.

How Did the OEM Structure the Supplier Data Flow?

The supplier data flow was structured around a tiered submission requirement that the OEM added to its supplier contracts in the prior year. Tier one suppliers — those providing major ECUs like the infotainment head unit, the telematics control unit, and the ADAS controller — were required to submit a CycloneDX SBOM with every software release. Tier two suppliers providing software components to tier ones were encouraged but not contractually required to submit SBOMs; the contractual flow-down depended on the relationship with the tier one.

Safeguard's supplier portal served as the ingestion point. Tier one suppliers uploaded SBOMs with their software delivery, or submitted them via API from their own CI systems. The platform validated each submission for structural conformance, completeness of component identifiers, and license field presence. Submissions that failed validation were rejected with specific remediation guidance.

For binary-only deliveries, Safeguard's binary composition analysis ran inside the OEM's environment to produce a derivative SBOM that supplemented whatever the supplier had provided. Discrepancies between the supplier's claimed SBOM and the derived SBOM were flagged for supplier follow-up.

How Did the Platform Handle the Reality of Embedded Automotive Software?

Embedded automotive software has characteristics that most enterprise SCA tools handle poorly. Components are often compiled for specific microcontrollers with limited symbol information. Binaries may be statically linked, stripped, and obfuscated. Real-time operating systems like AUTOSAR Classic do not have the same package ecosystem visibility as Linux. Safeguard's automotive-specific analyzers addressed several of these constraints.

For AUTOSAR Classic workloads, the platform's analyzer extracted component information from the standardized metadata files that the AUTOSAR toolchain produces. For AUTOSAR Adaptive and Linux-based ECUs (increasingly common in high-compute domains like ADAS and infotainment), the platform's standard container and binary analyzers applied. For proprietary RTOS environments, the OEM worked with Safeguard's solutions team to build specialized analyzers for the two most common RTOS platforms in the OEM's fleet.

The practical result was that across three vehicle platforms comprising approximately 440 ECUs, Safeguard could produce normalized composition data for roughly 88% of ECUs at the completion of the first integration cycle. The remaining 12%, concentrated in a few legacy supplier relationships, were on a roadmap for the subsequent cycle.

What Did the Vulnerability Monitoring Workflow Look Like?

The vulnerability monitoring workflow operated continuously across the ingested composition data. When a new CVE was disclosed that affected a component known to be present in the OEM's software inventory, Safeguard generated an alert that was routed to two destinations. The first was the OEM's internal cybersecurity operations team, which owned the operational response. The second was the supplier's designated contact, through an automated notification that included the affected component, the vehicle platforms where it was deployed, and the OEM's requested response timeline.

Response timelines were tied to the OEM's internal severity classification, which mapped CVSS severity to vehicle-system impact using the TARA methodology. A critical finding in a safety-relevant ECU triggered a 48-hour supplier acknowledgment requirement. A medium finding in a non-safety-relevant ECU triggered a 14-day acknowledgment requirement. The supplier portal tracked these SLAs, and the OEM's supplier management team used the tracking data in quarterly business reviews.

How Did This Fit Into the OEM's TARA Process?

The platform fed into the TARA process by providing the component-level vulnerability awareness that TARA assumed but previously could not reliably source. When a cybersecurity engineer performed a risk assessment on a new vehicle feature, they could query Safeguard for the current vulnerability profile of the involved ECUs and use that profile as input to the threat analysis. Before the platform was deployed, this query required manual outreach to supplier security contacts, which was slow and inconsistent.

The Head of Vehicle Cybersecurity noted that TARA throughput improved measurably after the platform rollout. The team completed 40% more TARA cycles in the first six months after deployment than in the comparable prior period, with no increase in headcount. The improvement was attributed primarily to the elimination of the manual supplier outreach step.

What Changed for Regulatory Homologation?

Regulatory homologation — the type approval process required to sell vehicles in most markets — increasingly includes cybersecurity assessments under UNECE R155 and related frameworks. The OEM's homologation team had historically assembled cybersecurity evidence for type approval through a multi-week internal effort that pulled data from multiple sources.

With Safeguard in place, the platform served as the authoritative source for the software supply chain portion of the homologation evidence package. Exports tied to specific vehicle platforms could be generated on demand and included the SBOM coverage, vulnerability posture, and remediation history that regulators expected. The homologation team reported that the time required to assemble the software supply chain evidence for a new vehicle platform approval decreased from approximately four weeks to under a week.

What Was the Impact on Supplier Relationships?

The impact was mixed in predictable ways. Suppliers with mature security practices welcomed the structured submission process because it provided them with a clear, repeatable way to meet the OEM's expectations. Suppliers without mature practices initially pushed back, citing cost and capability constraints. The OEM's supplier management team used the Safeguard portal's submission tracking to have data-driven conversations with lagging suppliers, and over the first year, the submission rate from tier one suppliers rose from approximately 60% to over 95%.

Two tier one suppliers were placed on improvement plans after sustained submission quality issues. One of these returned to acceptable performance within a quarter. The other was placed on a watch list for future program award decisions. The OEM's Head of Supplier Quality noted that the Safeguard data made supplier cybersecurity performance a tangible factor in procurement decisions for the first time.

How Safeguard.sh Helps

Safeguard.sh provides an automotive-tuned capability set covering the software supply chain requirements of ISO/SAE 21434, UNECE R155, and related frameworks. The platform supports AUTOSAR Classic and Adaptive workloads, Linux-based ECU software, and the RTOS environments most common in automotive embedded systems. For OEMs managing large and diverse supplier ecosystems, Safeguard provides a supplier portal, automated SBOM validation, binary composition analysis for binary-only deliveries, and vulnerability notification workflows with contractual SLA tracking. The maturity trajectory described here — moving from manual supplier outreach to a systematic supply chain security operation — is typical for OEM engagements that sustain executive sponsorship through the first year of rollout.

Never miss an update

Weekly insights on software supply chain security, delivered to your inbox.