Artifact GuideGLOBAL

ETSI EN 303 645 Requirements

Provision-by-provision requirements mapping for consumer IoT security (5.1-5.13).

Use this page to turn ETSI EN 303 645 into engineering controls, acceptance criteria, and evidence that survives audits and product iterations.

Author
Sorena AI
Published
Mar 4, 2026
Updated
Mar 4, 2026
Sections
14

Structured answer sets in this page tree.

Primary sources
2

Cited legal and guidance references.

Publication metadata
Sorena AI
Published Mar 4, 2026
Updated Mar 4, 2026
Overview

ETSI EN 303 645 is easiest to operationalize when you treat each provision as a testable claim: (1) the control exists, (2) it works for your product and associated services, and (3) you can prove it with repeatable evidence. The sections below translate the baseline requirements into practical implementation and evidence expectations.

Section 1

How to operationalize ETSI EN 303 645

Treat each provision as a control statement you can verify. For product teams, this typically means: a design decision, an implementation pattern, a test plan (conceptual + functional), and an evidence artifact that stays current per release.

When the standard says 'shall', plan for a durable control and objective evidence. When it says 'should', treat it as a risk-based requirement you either implement or explicitly justify with documented rationale and compensating controls.

  • Write a per-provision control narrative: intent -> design -> implementation -> verification -> evidence
  • Use release gates: security review, update signing verification, interface hardening checks, and regression tests
  • Store evidence like code, not like PDFs: versioned policies, versioned test results, and traceable tickets
Section 2

Provision 5.1 - No universal default passwords and strong authentication

The standard requires that device passwords (once not in factory default) are unique per device or user-defined, and it sets expectations for password generation and cryptographic best practices for authentication mechanisms.

This is where IoT security programs fail in practice: account provisioning, manufacturing flows, and support tooling must all align so that 'unique per device' is true in the field-not just in a requirement doc.

  • Eliminate shared default passwords across a product class; use per-device secrets or force user-set credentials
  • Generate pre-installed credentials using a mechanism that reduces automated attacks across a device class
  • Define authentication crypto and key management as a security parameter (owned, rotated, monitored)
Section 3

Provision 5.2 - Vulnerability disclosure policy (VDP) and timely remediation

You must publish a vulnerability disclosure policy with contact information and defined timelines for acknowledgement and status updates until resolution. The standard also sets expectations for timely action on disclosed vulnerabilities.

Operationally, this means you need an intake channel, triage workflow, remediation SLAs, and a way to communicate status updates to reporters and affected stakeholders.

  • Publish a VDP with a clear reporting channel and explicit acknowledgement + status-update timelines
  • Run a repeatable triage process (severity, exploitability, affected versions, mitigation options)
  • Maintain a vulnerability remediation playbook with release, communication, and rollout steps
Section 4

Provision 5.3 - Secure software updates and transparent support periods

ETSI EN 303 645 sets detailed expectations for updateability, secure installation, usability of updates, automatic updates, update checks, cryptography, authenticity and integrity verification, user notifications, and publishing a defined support period.

Build the update system as a security control: it must resist misuse, prevent downgrade attacks, verify trust relationships, and deliver timely security patches-even when supply-chain dependencies exist.

  • Implement a secure update mechanism (including anti-rollback) and make updates simple to apply
  • Use best-practice cryptography and verify authenticity + integrity of every update (device or trusted peer)
  • Publish a defined support period and communicate what happens at end-of-support (especially for constrained devices)
Section 5

Provision 5.4 - Secure storage of sensitive security parameters

Sensitive security parameters stored persistently must be stored securely. The standard also sets expectations for resisting tampering when a hard-coded per-device identity is used for security purposes.

In practice, this provision ties together secure storage, manufacturing provisioning, device identity, and incident response: if secrets are extractable, every other control becomes brittle.

  • Classify 'critical security parameters' (keys, tokens, secrets) and store them with platform protections
  • Protect device identity material against physical, electrical, and software tampering where applicable
  • Document provisioning and key lifecycle: generation, injection, backup/recovery, rotation, and revocation
Section 6

Provision 5.5 - Secure communications for critical security parameters

The standard addresses securing communications of critical security parameters: encrypt in transit where appropriate and protect confidentiality over remotely accessible interfaces, with secure management processes around those parameters.

This is not 'turn on TLS and forget it'. You need a threat model for your interfaces, certificate/key handling, and a plan for migration when crypto or endpoints change.

  • Encrypt critical security parameters in transit using cryptography appropriate to risk and technology
  • Harden remotely accessible interfaces so critical security parameters are never exposed in plaintext
  • Operate secure parameter management processes (issuance, access control, rotation, and incident handling)
Section 7

Provision 5.6 - Minimize attack surface (interfaces, debug, and information disclosure)

ETSI EN 303 645 requires disabling unused network and logical interfaces and minimizing unauthenticated disclosure of security-relevant information. It also addresses physical interface exposure and disabling debug interfaces when physically accessible.

This provision is where you win cheap security: if you ship less surface area, you have less to patch and less to explain in audits.

  • Disable unused network services, ports, protocols, and APIs in production builds
  • Avoid unauthenticated banners, version leakage, and verbose error messages on network interfaces
  • Disable or lock down debug interfaces (JTAG/UART) in software when physically accessible
Section 8

Provision 5.7 - Integrity of software and system operation

The integrity provisions focus on preventing unauthorized modification and preserving a trustworthy runtime state. For IoT devices, this often translates into secure boot, signed firmware, and protection of critical runtime components.

Integrity is also an evidence problem: you need to show how integrity is enforced and how integrity failures are detected and handled.

  • Implement signed firmware and verify before execution where your platform supports it
  • Define how integrity violations are handled (reject, fail-safe, report, recover)
  • Keep integrity evidence: signing keys policy, build pipeline attestations, verification logs
Section 9

Provision 5.8 - Personal data confidentiality and user transparency

The standard sets expectations for protecting the confidentiality of personal data in transit (especially sensitive personal data) and for documenting external sensing capabilities in a clear, transparent, and accessible way for the user.

A strong implementation links privacy engineering to product UX: data is protected, and the user can understand what the device can sense and transmit.

  • Protect personal data in transit with best-practice cryptography; apply stronger guarantees for sensitive personal data
  • Document external sensing capabilities in user-facing language (what is sensed, when, and why)
  • Align privacy controls with update and vulnerability processes (privacy bugs also need CVD and fixes)
Section 10

Provision 5.9 - Resilience to outages and clean recovery

ETSI EN 303 645 recommends building resilience into devices and services, including local functionality during network outages and clean recovery after power loss.

Resilience is security-relevant: unsafe failure modes become exploitable. Treat outage behavior, state recovery, and 'expected stable state' as part of your security acceptance criteria.

  • Define safe degraded modes for loss of network and loss of power
  • Implement clean recovery and orderly reconnection behavior
  • Test outage scenarios and capture evidence (power-cycle tests, network partition tests)
Section 11

Provision 5.10 - Telemetry and security anomaly detection

If telemetry is collected, it should be examined for security anomalies. This pushes teams to treat telemetry as a security sensor, not only as a product analytics pipeline.

Telemetry is also a trust boundary: collect minimally, protect it, and ensure anomaly detection does not leak sensitive information.

  • Define telemetry security signals: update failures, auth failures, integrity violations, interface scans
  • Create anomaly detection and response playbooks (triage, containment, remediation)
  • Retain evidence: alert rules, incident tickets, and post-incident corrective actions
Section 12

Provision 5.11 - User data deletion and clear instructions

The standard requires simple deletion of user data from the device and expectations for removing personal data from associated services, with clear instructions and confirmation of deletion where applicable.

Deletion is a lifecycle control: it must work on-device, in associated services, and across customer support flows (returns, resale, transfers).

  • Provide simple, reliable user data erasure on the device (reset that actually deletes)
  • Support deletion from associated services and provide user instructions and confirmations
  • Test reset flows across firmware versions and associated app/service versions
Section 13

Provision 5.12 - Secure installation and maintenance with usable security

Installation and maintenance should involve minimal user decisions and follow security best practice on usability. The manufacturer should provide guidance on secure setup and how to check whether the device is securely set up.

Usable security is a compliance multiplier: if secure setup is hard, the field reality will violate your intended controls.

  • Default to secure settings and minimize risky configuration choices during onboarding
  • Provide step-by-step secure setup guidance and a 'security status' check for users
  • Ensure guidance stays accurate across app versions, firmware versions, and SKUs
Section 14

Provision 5.13 - Input validation for UIs, APIs, and service-to-device flows

The standard requires validating data input via user interfaces, APIs, and between networks in services and devices. This is the foundation for preventing common classes of vulnerabilities in IoT ecosystems.

Make input validation measurable: schema validation, size and type limits, authentication and authorization checks, and security testing for all externally reachable surfaces.

  • Validate and sanitize inputs across device UIs, network APIs, and service integrations
  • Add security tests: fuzzing, boundary tests, authZ tests, and regression tests for prior vulnerabilities
  • Log validation failures safely (no sensitive data leakage) and monitor for abuse patterns
Recommended next step

Turn ETSI EN 303 645 Requirements into an operational assessment

Assessment Autopilot can take ETSI EN 303 645 Requirements from turning the requirements into assigned actions to a reusable workflow inside Sorena. Teams working on ETSI EN 303 645 can keep owners, evidence, and next steps aligned without copying this guide into separate documents.

Primary sources

References and citations

Related guides

Explore more topics