What is SESIP and who publishes it?

SESIP (Security Evaluation Standard for IoT Platforms) is a cybersecurity evaluation methodology published by GlobalPlatform, an industry consortium headquartered in Redwood City. The public 1.0 release was issued in May 2020 under reference GP_FST_070. SESIP is derived from Common Criteria ISO/IEC 15408 but optimised for constrained IoT platforms, silicon components, embedded operating systems, application stacks and complete IoT products. It targets shorter evaluation timelines and reuse of results through composition along the IoT stack.

How many SESIP levels exist and what do they cover?

SESIP defines five assurance levels numbered 1 to 5, aligned with the Common Criteria AVA_VAN scale. Level 1 is a self-assessment bundled with AVA_VAN.1 (review of public vulnerabilities). Level 2 adds AVA_VAN.2 and the involvement of an accredited laboratory. Level 3 introduces AVA_VAN.3 and a black-box penetration test. Level 4 requires AVA_VAN.4 with side-channel analysis and fault injection. Level 5 targets state-of-the-art attack resistance (AVA_VAN.5), typically for secure elements and eUICC.

How is SESIP different from classical Common Criteria?

SESIP reuses the Common Criteria vocabulary (Security Target, Security Functional Requirements, AVA_VAN) but simplifies the process. Common Criteria require a tailored Protection Profile and an EAL evaluation covering all of the ADV, AGD, ALC, ATE and AVA classes. SESIP defines a single IoT-oriented profile, reduces documentation to the evidence strictly needed, and structures assurance packages for composition (a SESIP-evaluated component can be integrated into a SESIP-evaluated platform or product without full re-verification).

Is SESIP recognised by European regulation?

Yes, indirectly. The EN 18031 series of harmonised standards for RED article 3.3 allows, at the higher assurance levels, the use of a SESIP evaluation as evidence of conformity. The European standard ETSI EN 17927, published in 2023, derives directly from SESIP for European IoT cybersecurity evaluation. The future EUCC scheme, set under the Cybersecurity Act and aligned with the CRA, likewise foresees the use of SESIP as a recognised methodology for components and platforms.

What is the relationship between SESIP and PSA Certified?

PSA Certified, an operational programme led by Arm and its partners (Brightsight, Riscure, SGS, TrustCB, UL), aligned PSA level 2 and PSA level 3 with SESIP profiles. PSA level 2 maps in practice to SESIP level 3 (black-box penetration test), PSA level 3 to SESIP level 4 or 5 depending on scope. A PSA Certified level 2 or 3 certificate issued from 2022 onwards explicitly names the SESIP profile applied, which facilitates composition into a final product evaluation.

How long is a SESIP certificate valid?

The default validity period of a SESIP certificate is set by the scheme under which the evaluation is conducted. Current practice is a five-year validity, conditional on a vulnerability management programme maintained by the manufacturer. Discovery of a critical vulnerability on the product or on a reused component can suspend the certificate. A major change to the target (chip swap, firmware update affecting attack surface) triggers at least a partial re-evaluation.

Is SESIP suitable for a standalone microcontroller?

Yes. SESIP is explicitly designed to allow evaluation at three levels of granularity, component (chip, eUICC, secure element), platform (chip plus secure OS plus crypto stack), and product (platform integrated into a final IoT device). The evaluation of a standalone microcontroller yields a component certificate reusable by downstream integrators, with a documented composition rationale. This modularity is the main contribution of SESIP over traditional Common Criteria evaluations.

What are the most common pitfalls in a SESIP project?

Three recurring pitfalls. First, an under-scoped Security Target, where the manufacturer poorly describes the operational environment and assumptions, weakening the evaluation. Second, confusion between SESIP levels and Common Criteria EAL, SESIP level 4 is not equivalent to EAL 4, the scales are different. Third, badly justified composition, reusing a component certificate without formally demonstrating that the certificate assumptions hold in the integration context leads to a refusal by the laboratory or certification body.

SESIP: IoT platform security evaluation methodology

Author Hugues Orgitello Last updated 27 May 2026 ~16 min read

Guide. SESIP / GlobalPlatform

SESIP (Security Evaluation Standard for IoT Platforms) is the cybersecurity evaluation methodology published by GlobalPlatform in May 2020 under reference GP_FST_070. Derived from Common Criteria ISO/IEC 15408 but optimised for constrained IoT platforms, it organises evaluation into five assurance levels aligned with the AVA_VAN scale. Its distinctive feature is composition: an evaluated component (chip, secure OS, crypto stack) can be integrated into a platform and then a final product without full re-verification. SESIP is the methodology recognised by the EN 18031 series at the higher assurance levels, and the technical backbone of ETSI EN 17927 and the future EUCC scheme. This guide sets out its origin, evaluation structure, articulation with PSA Certified and Common Criteria, and the most frequent pitfalls.

Origin and the gap addressed

Until the late 2010s, formal cybersecurity evaluation of electronic components was largely conducted under the Common Criteria Recognition Arrangement (CCRA). Common Criteria (CC), formalised in ISO/IEC 15408 and the supporting methodology ISO/IEC 18045, remain the worldwide reference for high-assurance evaluations, in particular in the banking, governmental and identity domains.

Three characteristics of CC make them ill-suited to the mass-market IoT segment:

typical durations of an EAL 4 or higher evaluation (twelve to twenty-four months) are incompatible with the time-to-market of a connected sensor or a smart home device;
the cost of a full CC evaluation (in the order of several hundred thousand euros for a non-trivial target) does not align with the unit economics of a mid-volume IoT product;
the absence of native composition machinery forces full re-evaluation of the stack on every product iteration, whereas IoT relies precisely on the standardised stacking of bricks (silicon, OS, middleware, application).

GlobalPlatform, the industry consortium already responsible for the Card Specification and TEE (Trusted Execution Environment) specifications, started the SESIP project in 2018 to close this gap. Public release 1.0 was issued on 26 May 2020 under reference GP_FST_070, subtitled "Security Evaluation Standard for IoT Platforms". An incremental revision is in preparation, with alignment expected against the 2022 edition of ISO/IEC 15408 (the major CC revision).

SESIP inherits the CC vocabulary: Security Target, Security Problem Definition, Security Objectives, Security Functional Requirements, AVA_VAN class. It introduces, in contrast, a single IoT-oriented profile (rather than the bespoke Protection Profile of CC), a fixed and reduced set of assurance classes, and an explicit certificate composition mechanism.

Three levels of evaluation granularity

SESIP formally distinguishes three target-of-evaluation granularities, matching the three typical layers of an IoT stack.

Target level	Description	Typical examples
Component	Atomic hardware or software block of the stack	MCU with secure boot, eUICC, secure element, trusted OS
Platform	Hardware and software assembly providing security services	MCU + bootloader + secure OS + crypto library
Product	Complete IoT device integrating a platform	Smart meter, smart-home gateway, industrial sensor

This granularity has a direct economic consequence: a silicon supplier can have its component certified once, then deliver the certificate and supporting documentation to all downstream integrators. The integrator reuses that certificate in the evaluation of its platform or product, demonstrating through a composition argument that the upstream certificate assumptions still hold in the integration context.

The five SESIP assurance levels

SESIP defines five assurance levels numbered 1 to 5, aligned with the AVA_VAN assurance components of Common Criteria. AVA_VAN is the "Vulnerability Analysis" class of CC, graded across five levels by the attack potential resisted.

SESIP level	Equivalent AVA_VAN	Type of evaluation	Laboratory involvement	Expected resistance
SESIP 1	AVA_VAN.1	Guided self-assessment	Optional	Known public vulnerabilities
SESIP 2	AVA_VAN.2	Accredited laboratory evaluation	Mandatory	Attacker with basic potential
SESIP 3	AVA_VAN.3	Laboratory plus black-box penetration test	Mandatory	Attacker with enhanced potential
SESIP 4	AVA_VAN.4	SESIP 3 plus side-channel analysis and fault injection	Mandatory, specialised equipment	Attacker with moderate potential, moderate equipment
SESIP 5	AVA_VAN.5	Resistance to state-of-the-art attacks	Mandatory, advanced equipment	Attacker with high potential, high equipment and expertise

The AVA_VAN scale is documented in the Common Methodology for Information Technology Security Evaluation (CEM, ISO/IEC 18045). Attack potential is estimated through five factors: preparation time, expertise required, knowledge of the target, window of opportunity, hardware equipment required. An aggregate score determines the AVA_VAN level the target resists.

SESIP 1: tool-supported self-assessment

Level 1 covers documentary conformity and review of public vulnerabilities. The manufacturer produces the Security Target itself, describes the security measures, and provides evidence that known public vulnerabilities (CVE, CWE) applicable to its product are handled. No laboratory intervention is mandatory. The evaluation is reviewed by the certification body on the documentation alone.

SESIP 2: accredited laboratory, functional testing

Level 2 introduces examination by a laboratory accredited under a GlobalPlatform scheme. The laboratory verifies the coherence of the Security Target, exercises the declared security functions, and conducts an extended vulnerability review. Expected resistance covers an attacker with basic potential: standard tools, no specialised equipment, no in-depth knowledge of the target.

SESIP 3: black-box penetration testing

Level 3 adds a logical black-box penetration test. The laboratory actively attempts to bypass the security functions through interfaces accessible to the attacker, without access to source code or keys. Expected resistance covers an attacker with enhanced potential: non-trivial expertise, advanced software tooling, but without sophisticated hardware equipment.

SESIP 4: side-channel and fault injection

Level 4 introduces two families of hardware tests:

side-channel analysis, exploiting information leaks through power consumption measurement (Simple Power Analysis, Differential Power Analysis), electromagnetic emission or timing response to extract cryptographic secrets;
fault injection analysis, deliberately injecting disturbances (voltage glitches, electromagnetic pulses, focused laser) to deviate execution and bypass security checks.

These tests require specialised laboratory equipment (high-end oscilloscopes, EM benches, laser benches, micro-positioning stations) and deep expertise. Level 4 is typically targeted by payment components, high-end eUICC and commercial secure elements.

SESIP 5: state-of-the-art resistance

Level 5 targets resistance to state-of-the-art attacks worldwide: very high attacker potential, extended preparation, advanced equipment (scanning electron microscope, FIB for circuit editing, substrate analysis). This level is typical of governmental identity components, high-end banking cards and certified hardware security modules (HSM). Very few mass-market IoT products target this level, where evaluation cost runs in the hundreds of thousands of euros.

SESIP, Common Criteria and PSA Certified compared

The table below maps the three cybersecurity evaluation frameworks most used for IoT components and platforms.

Dimension	SESIP	Common Criteria (ISO/IEC 15408)	PSA Certified
Publisher	GlobalPlatform	ISO, CCRA, national schemes	Arm and partners
Reference	GP_FST_070 v1.0 (May 2020)	ISO/IEC 15408:2022 and CEM	PSA Certified specifications
Target	IoT platforms and components	All IT products	IoT platforms and Arm MCUs
Profile	Single SESIP profile	Bespoke Protection Profile	PSA profile referencing SESIP
Levels	5 (SESIP 1 to 5)	7 EAL plus AVA_VAN.1 to .5	3 (PSA level 1, 2, 3)
Native composition	Yes, explicit mechanism	CC composition (rare, heavy)	Yes, inherited from SESIP
Typical evaluation duration	Weeks to months	Six to twenty-four months	Weeks to months
Typical evaluation cost	Tens to hundreds of thousands of euros	Hundreds of thousands of euros	Tens to hundreds of thousands of euros
Regulatory recognition	EN 18031 (EU), EUCC (in progress)	CCRA, EUCC, national schemes	Recognised via SESIP by EN 18031
Typical use case	IoT MCU, eUICC, secure OS, IoT device	Smart card, HSM, governmental OS	Arm Cortex-M, Cortex-A MCU and platforms

The rough equivalence between SESIP and PSA Certified levels is the following: PSA level 1 is a self-attestation aligned with the PSA Functional API 10-Question Assessment, with no direct SESIP mapping; PSA level 2 maps in practice to SESIP level 3 (laboratory plus penetration testing); PSA level 3 maps to SESIP level 4 (side-channel and fault injection), with possible extension to SESIP level 5 for the most demanding targets. This mapping is not absolute and depends on the Security Target scope.

The equivalence between SESIP levels and Common Criteria EAL levels does not exist formally. SESIP level 4 is not equivalent to EAL 4. EAL levels aggregate assurance across seven classes (ADV, AGD, ALC, ATE, AVA, and formerly ACM, ASE, APE), whereas SESIP levels concentrate on the AVA_VAN scale. Conflating the two scales in a declaration of conformity is a common pitfall and a frequent ground for rejection by certification bodies.

SESIP profile and composition-based evaluation

The SESIP profile plays the role that a Protection Profile (PP) plays in Common Criteria: it sets the frame for the Security Target. Where a CC PP is specific to the product or product family considered, the SESIP profile is unique and generic, applicable to any IoT platform. The Security Target instantiates the profile for a given product by filling out the mandatory sections:

TOE description (Target Of Evaluation), what is evaluated, what is not, the precise boundary;
Security problem definition, threats considered, environmental assumptions, organisational policies;
Security objectives, security objectives covered by the product and by the environment;
Security functional requirements (SFR), functional requirements selected from the SESIP catalogue (which carries over a subset of ISO/IEC 15408-2);
TOE security functionality (TSF), concrete description of the mechanisms implemented.

Composition of certificates

The composition mechanism is the central contribution of SESIP. Consider a component evaluated at SESIP level X with certificate C1. Consider a platform integrating this component and evaluated at SESIP level Y (with Y at most X). The platform evaluation reuses C1 as evidence, without fully re-verifying the component, under two conditions:

the environmental assumptions of the component (operating environment assumptions) must be satisfied in the platform integration context;
the security policies declared by the component must be coherent with those of the platform.

These two conditions are documented in a composition rationale added to the platform Security Target. The laboratory evaluating the platform verifies that rationale but does not re-run tests already executed on the component.

Example: silicon to platform to product composition

The example below illustrates typical composition along an IoT stack.

Layer	Target of evaluation	Target level	Reused element
Silicon	MCU with secure boot and secure storage	SESIP 4 (side-channel)	Component certificate, chip Security Target
Platform	MCU + bootloader + secure OS + crypto stack	SESIP 3	MCU certificate + platform Security Target
Product	IoT sensor integrating the platform + application firmware	SESIP 2	Platform certificate + product Security Target

A final manufacturer integrating a SESIP 4 certified MCU into its product can target SESIP 2 on the complete product by leveraging the upstream certificate. The evaluation effort focuses on the application firmware and integration, not on the MCU itself.

This logic transforms the economics of IoT certification: an initial effort concentrated at the silicon supplier is monetised across all downstream products. A final product not certified today can target a modest SESIP level with reasonable effort, provided that it integrates components already certified at SESIP level equal or higher.

SESIP in the European landscape: EN 17927, EN 18031, EUCC

SESIP entered the European regulatory framework through three distinct channels.

ETSI EN 17927

Published in 2023, ETSI EN 17927 ("Methods and protocols, security evaluation of IoT") is a European standard derived directly from SESIP. The text reproduces the five-level structure, the composition mechanism and the functional requirements catalogue. EN 17927 acts as the European normative counterpart to SESIP, usable in EU regulatory schemes.

EN 18031 series

The EN 18031-1/-2/-3 series, published in August 2024 and listed in the Official Journal of the European Union in January 2025, defines harmonised standards for conformity to RED article 3.3 (cybersecurity of radio equipment). EN 18031 distinguishes several assurance levels and, at the higher levels, allows the use of a SESIP level 3 or higher evaluation as evidence of conformity. In practice, a radio IoT device targeted by RED 3.3 can leverage the SESIP certificate of its platform to demonstrate compliance with essential requirements, alongside the documentation foreseen by EN 18031.

For the wider European radio placing-on-the-market context, see our ETSI EN 303 645 guide, which covers the complementary voluntary baseline for consumer IoT.

EUCC scheme under the Cybersecurity Act

The European Common Criteria scheme (EUCC) is the first certification scheme adopted under the Cybersecurity Act (Regulation (EU) 2019/881). Adopted by Implementing Regulation (EU) 2024/482 and applicable since February 2025, EUCC carries over the Common Criteria structure and provides for SESIP incorporation in IoT evaluations.

The connection between EUCC, SESIP and the Cyber Resilience Act (CRA, Regulation (EU) 2024/2847, applicable from 11 December 2027) is still being shaped by ENISA. The current outline foresees that for "important" and "critical" products under the CRA, certification under EUCC or a recognised programme (including SESIP) will serve as presumption of conformity. The exact timeline and scope depend on implementing acts expected in 2026 and 2027.

SESIP and PSA Certified: an operational alignment

PSA Certified is the operational certification programme led by Arm and a consortium of laboratories (Brightsight, Riscure, SGS Brightsight, TrustCB, UL). Initially structured around three PSA-specific levels, the programme aligned in 2022 its PSA level 2 and PSA level 3 with SESIP profiles.

PSA level	Indicative SESIP mapping	Target type	Dominant method
PSA level 1	No direct SESIP mapping	Guided self-attestation via 10-Question Assessment	Documentary
PSA level 2	SESIP level 3	Chip plus PSA Root of Trust	Black-box penetration test
PSA level 3	SESIP level 4 (sometimes 5)	Chip for high-demand applications	Side-channel and fault injection

A PSA Certified level 2 or 3 certificate issued from 2022 onwards explicitly names the SESIP profile applied, allowing a downstream integrator to reuse that certificate in a SESIP composition. Conversely, a manufacturer aiming for PSA certification goes through a SESIP-accredited laboratory. PSA Certified and SESIP are therefore not competitors but complements: SESIP is the methodology, PSA an operational programme that instantiates it for the Arm ecosystem.

For non-Arm cases (RISC-V, MIPS, embedded x86), evaluation goes directly through SESIP without a PSA intermediate.

Accredited laboratory ecosystem

SESIP evaluation requires the involvement of a laboratory accredited by GlobalPlatform (except at level 1, where self-assessment is admitted). GlobalPlatform maintains a public list of accredited laboratories, updated periodically on the sesip.globalplatform.org portal. The main players at the time of writing include:

Brightsight (Netherlands, part of SGS), historically the reference laboratory for banking secure components and eUICC;
Riscure (Netherlands), recognised specialist in side-channel analysis and fault injection, particularly at level 4;
SGS (international group), broad portfolio covering IoT product certification;
TrustCB (Netherlands), ITSEF certification scheme and certificate-issuing body;
UL Solutions (United States and Europe), diversified laboratory covering the consumer and industrial IoT segments.

Several additional European laboratories (Serma Safety and Security in France, T-Systems in Germany, applus+ Laboratories in Spain) are accredited or under accreditation for SESIP levels 2 and 3. The geographic coverage of the SESIP scheme broadens as EN 17927 and EN 18031 gain weight.

Certificate life cycle and maintenance

A SESIP certificate is not a frozen deliverable. Its validity rests on three conditions sustained over time.

Initial duration

The default validity is set by the scheme under which the evaluation is conducted. Current practice observes an initial validity of five years for SESIP levels 1 to 3, sometimes reduced to two or three years at levels 4 and 5 where the target is more exposed to the rapid evolution of attack techniques.

Vulnerability maintenance

The manufacturer must maintain a vulnerability surveillance and handling programme covering the validity period. Any publicly disclosed vulnerability affecting the target or a reused component must be assessed for its impact on the certificate. A critical vulnerability that cannot be remediated may lead to suspension or withdrawal of the certificate by the certification body.

Re-evaluation on evolution

A major change to the target triggers at least a partial re-evaluation:

change of silicon chip or major silicon revision;
modification of the bootloader or the chain of trust;
firmware update adding functions accessible to the attacker;
introduction of new external interfaces (additional radio, debug port).

Minor changes (security patches, application adjustments without attack surface change) are usually handled through a certificate addendum, faster than a full re-evaluation.

Frequent pitfalls in a SESIP project

Field experience from the early years of the programme has surfaced several recurring pitfalls for manufacturers approaching SESIP. The methodology cannot be improvised, and rigorous up-front scoping avoids costly late refusals. AESTECHNO designs IoT products and tracks the maturity of these methodologies for cybersecurity architecture decisions.

Summary

SESIP is today the cybersecurity evaluation methodology best aligned with the industrial reality of IoT: it carries the rigour of Common Criteria, simplifies implementation, allows certificate composition along the silicon-platform-product stack, and benefits from growing European regulatory recognition via EN 17927, EN 18031 and the future EUCC. For an IoT manufacturer anticipating CRA obligations from 2027, familiarity with SESIP is no longer optional. Architecture decisions taken at design phase, in particular the selection of a silicon component already certified under SESIP, directly drive the cost and duration of the product evaluation. The follow-up of Cybersecurity Act and CRA implementing acts, expected in 2026 and 2027, will complete the regulatory picture and fix the precise place of SESIP in the European certification landscape.