Electronics Guide

Common Criteria Evaluation Equipment

Common Criteria is the most widely recognized international framework for evaluating information technology security. Its current release, CC:2022, is published as ISO/IEC 15408 parts 1 through 5, with the companion evaluation methodology in ISO/IEC 18045. Evaluations are performed by licensed laboratories, called Common Criteria Testing Laboratories in the United States and IT Security Evaluation Facilities in Europe, and the resulting certificate is issued by a national certification body rather than by the laboratory itself.

The seven Evaluation Assurance Levels, EAL1 through EAL7, remain in the standard, but they no longer determine international recognition on their own. Under the 2014 revision of the Common Criteria Recognition Arrangement, certificates are mutually recognized when the evaluation follows a collaborative Protection Profile written by an international Technical Community; evaluations against a vendor-authored Security Target are recognized only up to EAL2 augmented with flaw remediation. Higher assurance has historically been recognized regionally through the SOG-IS agreement in Europe. That role is passing to the EUCC, the European Common Criteria-based certification scheme established by Commission Implementing Regulation (EU) 2024/482, adopted on 31 January 2024 and applicable from 27 February 2025. EUCC anchors its "substantial" and "high" assurance levels to the vulnerability analysis family AVA_VAN rather than to the EAL number alone.

That shift matters for equipment planning. Assurance levels up to AVA_VAN.2 can be satisfied with functional test harnesses, coverage measurement tools, static and dynamic code analysis platforms, and protocol fuzzers. AVA_VAN.4 and AVA_VAN.5, required for smart cards, secure elements, and hardware security modules, oblige the evaluator to attempt real key-recovery attacks with a full side-channel and fault-injection bench. A laboratory that intends to evaluate a secure element must therefore own the same equipment an attacker would use, and must be able to show that it operated that equipment competently.

Beyond the attack bench, a Common Criteria facility needs infrastructure for handling the target of evaluation and its evidence. This includes isolated test networks that keep concurrent evaluations from contaminating one another, safes or access-controlled rooms for development-grade samples and source code, configuration management for every test script and firmware image used, and logging that reconstructs who touched what and when. Evaluation records are retained for years, because a certificate can be challenged or a product re-evaluated long after the original testing concludes.

FIPS 140-3 Validation Equipment

FIPS 140-3 establishes the security requirements for cryptographic modules used by the United States and Canadian federal governments and, by extension, across much of the regulated commercial market. Unlike its predecessor, FIPS 140-3 does not restate its requirements: it adopts ISO/IEC 19790:2012 as the security requirements and ISO/IEC 24759 as the derived test requirements, with a short set of national modifications. The Cryptographic Module Validation Program, operated jointly by NIST and the Canadian Centre for Cyber Security, began accepting FIPS 140-3 submissions in September 2020, stopped accepting new FIPS 140-2 submissions on 1 April 2022, and will move every remaining FIPS 140-2 certificate to the Historical List on 21 September 2026. Testing is performed by laboratories accredited under the NIST National Voluntary Laboratory Accreditation Program.

Validation covers four security levels. Level 1 requires only approved algorithms and basic documentation. Level 2 adds tamper evidence and role-based authentication. Level 3 adds tamper detection and response with zeroization of critical security parameters, along with identity-based authentication and separated interfaces for plaintext keys. Level 4 adds a tamper-detection envelope, protection against environmental attack, and resistance to non-invasive attack. Each step upward adds test equipment rather than merely paperwork.

Algorithm testing runs through the Cryptographic Algorithm Validation Program, which is now automated end to end by the Automated Cryptographic Validation Protocol. Rather than exchanging response files by hand, the laboratory runs an ACVP client that requests test vectors from the NIST server over an authenticated session, drives the module under test, and submits results for machine verification. The vector families include known-answer tests, Monte Carlo tests that chain thousands of iterations to expose state-handling defects, and multi-block message tests. Automation has cut algorithm testing from weeks to hours and removed a large class of transcription errors, though the laboratory still has to build and maintain the harness that connects the client to the module's cryptographic interface.

Entropy is validated separately. Since 2022 the CMVP has issued standalone certificates through the Entropy Source Validation program for noise sources conforming to NIST SP 800-90B. Validation requires the vendor to expose raw, unconditioned samples from the noise source, plus restart datasets that show the source recovers unpredictably after power cycling. Laboratories run the SP 800-90B estimators over these datasets to compute a min-entropy figure, in either the IID or the non-IID track, and verify the health tests, startup tests, and continuous tests that the module performs on itself. The deterministic random bit generator built on top of the source is tested against SP 800-90A, and the overall construction against SP 800-90C.

Physical and environmental testing occupies the rest of the laboratory. Thermal chambers and programmable supplies verify environmental failure protection or environmental failure testing at Level 4, confirming that the module either continues operating correctly or zeroizes its keys when temperature or voltage leaves the defined range. Tamper-response testing exercises intrusion switches, mesh sensors, and secure erasure paths, and instruments the response so that zeroization is shown to complete before an attacker can reach the protected data. Non-invasive attack mitigation at Levels 3 and 4 is assessed by the methods of ISO/IEC 17825, described below.

EMV Type Approval Platforms

EMVCo maintains the chip-payment specifications and the type approval process that keeps cards and terminals interoperable worldwide. The organization takes its name from the original Europay, Mastercard, and Visa collaboration of the 1990s; it is now owned in equal shares by six payment networks: American Express, Discover, JCB, Mastercard, UnionPay, and Visa. Type approval is granted by EMVCo on the basis of testing performed at qualified laboratories, and it is organized into two levels that require quite different equipment.

Level 1 approval covers the electromechanical and radio-frequency interface, independently of any payment application. Contact interfaces are tested against the electrical and transport-protocol behavior derived from ISO/IEC 7816, checking supply voltages, activation and deactivation sequences, clock and timing tolerances, and character-level protocol handling. Contactless interfaces are tested against the EMV Contactless Communication Protocol, which layers on ISO/IEC 14443 Types A and B, checking operating field strength, modulation depth and waveform shape, load modulation amplitude, frame timing, and anticollision behavior. This testing depends on calibrated reference devices: a reference card that a terminal must drive correctly, and a reference reader that a card must respond to correctly, positioned by mechanical jigs so that coupling conditions are reproducible. Commercial benches from suppliers such as Comprion, Keolabs, and Micropross implement these reference conditions, and laboratories keep their analog test setups under regular recalibration because small differences in loading or antenna geometry can shift a marginal device from pass to fail.

Level 2 approval covers the kernel, the software in the terminal that runs the payment transaction: application selection, offline data authentication, cardholder verification, terminal risk management, and the online authorization exchange. Contactless kernels are specified as a family, C-1 through C-8, each aligned with a payment brand's transaction flow. Level 2 testing is driven by card simulators and automated test suites that replay thousands of scripted scenarios, including malformed responses, interrupted transactions, and deliberately hostile card behavior. Protocol analyzers capture the resulting exchanges for evidence. Certification beyond Level 2, sometimes called Level 3 or end-to-end certification, is performed by acquirers and individual payment brands rather than by EMVCo, and it validates the specific terminal, application, and host combination in production conditions.

Chip security is handled by a separate track. The EMVCo Security Evaluation Process assesses integrated circuits and card platforms at accredited security evaluation laboratories using attack-potential scoring closely related to the Common Criteria smart-card methodology. Those laboratories operate the same side-channel, fault-injection, and invasive-analysis equipment described later in this article, and their findings feed the composite evaluations that let a card product inherit assurance from a certified chip.

PCI PTS Device Approval

PIN Transaction Security is the PCI Security Standards Council program that approves payment hardware, and it is the payment scheme most dependent on physical test equipment. The Point of Interaction requirements apply to PIN entry devices, unattended terminals, and card readers; companion standards cover hardware security modules used in payment processing and the secure environments used for card production. Approved devices are published on a public list with expiry dates, so approval is a recurring obligation rather than a one-time event. The requirements are issued as a modular document, currently in a major revision from the version 6 series to version 7.0, grouped into evaluation modules covering core physical and logical security, terminal integration, communications and interfaces, and life-cycle security. A device is assessed only against the modules that match its architecture.

Physical evaluation is adversarial and destructive. Laboratories attempt to penetrate the tamper-responsive enclosure by drilling, milling, and cutting under microscopy, to probe the keypad matrix or the secure processor without triggering the tamper mesh, and to defeat the device by cooling, heating, or varying the supply. Testing also targets observation attacks against PIN entry, including camera placement against the keypad shield and acoustic or electromagnetic emanations from key presses. Every successful or partially successful attack is costed in the same way a Common Criteria evaluation costs one: the time, expertise, equipment, and device-specific knowledge required to identify the attack, and then to exploit it on further units. A device passes when the cheapest viable attack exceeds the threshold the standard sets.

Logical evaluation examines firmware authentication, key loading and key management, the account data encryption function known as secure reading and exchange of data, and the open protocol stack exposed to the network. Because payment terminals now carry general-purpose operating systems, touchscreens, and cellular or Wi-Fi radios, the logical portion of a PTS evaluation increasingly resembles a Common Criteria vulnerability analysis, with firmware extraction, static analysis, and fuzzing of the exposed interfaces.

PCI DSS Compliance Testing

The PCI Data Security Standard sits at the other end of the spectrum from device approval. It governs the organizations that store, process, or transmit cardholder data, and it is assessed against an operating environment by a Qualified Security Assessor rather than against a product in a laboratory. Version 4.0 was published in 2022, version 3.2.1 was retired on 31 March 2024, version 4.0.1 followed in 2024, and the future-dated requirements introduced by version 4.0 became mandatory on 31 March 2025. The standard retains twelve principal requirements covering network security controls, secure configuration, data protection, access control, monitoring, and policy.

Technical validation relies on scanning and analysis tooling rather than instruments. External vulnerability scans must be performed quarterly by an Approved Scanning Vendor; internal scans and authenticated configuration assessments are run by the entity itself. Segmentation testing verifies that the cardholder data environment is genuinely isolated from the rest of the network, which in practice means attempting to route traffic across the boundary from every adjacent segment. Wireless scanning identifies unauthorized access points. Annual penetration testing, at both the network and the application layer, must follow a documented methodology and demonstrate that segmentation controls hold.

Application security is covered separately by the PCI Software Security Framework, whose Secure Software Standard and Secure Software Lifecycle Standard replaced the retired Payment Application Data Security Standard. Assessment uses web and API scanners, interactive and static analysis platforms, and manual code review focused on injection, authentication, session handling, and cryptographic misuse. Logging and monitoring requirements drive the remaining tooling: log collection with tamper-evident storage, cryptographic integrity protection of audit trails, file integrity monitoring, and correlation in a security information and event management platform. Evidence from these systems is what an assessor examines, so retention and integrity matter as much as detection quality.

Side-Channel Evaluation Equipment

Side-channel analysis recovers secrets from physical quantities that leak while a device computes: instantaneous power draw, electromagnetic emission, execution time, and in some cases sound or photonic emission. A cryptographic implementation can be mathematically sound and still surrender its key in minutes to a well-instrumented bench, which is why side-channel resistance has become a mandatory element of high-assurance certification rather than an optional extra.

A power analysis bench is built around a high-bandwidth oscilloscope, typically sampling at one to ten gigasamples per second with eight to twelve bits of vertical resolution, capturing the voltage across a small shunt resistor in the supply return or the output of a low-noise current probe. A trigger derived from the device's own I/O or a dedicated synchronization pin aligns each trace to the start of the cryptographic operation. Analysis software then applies differential power analysis, correlation power analysis against a leakage model such as Hamming weight or Hamming distance, or template attacks built from a profiling device. Trace alignment, filtering, and resynchronization tools compensate for clock jitter and randomized delays inserted as countermeasures.

Electromagnetic analysis substitutes near-field magnetic or electric probes for the current sensor. Because a probe with a loop of a few hundred micrometers can be positioned over a specific region of a die, EM analysis often defeats countermeasures that survive whole-device power analysis: it isolates the leakage of one functional block from the noise of the rest of the chip. Motorized XYZ probe stations scan the die surface systematically to build leakage maps, and low-noise amplifiers and spectrum analyzers extend the usable measurement range.

Two evaluation philosophies coexist. Conformance testing under ISO/IEC 17825 uses the Test Vector Leakage Assessment framework, comparing sets of traces gathered with fixed and with random inputs using Welch's t-test and flagging leakage when the statistic exceeds a threshold of 4.5. The 2024 revision of the standard extends TVLA to higher-order detection for masked implementations and offers guidance for asymmetric and post-quantum algorithms. TVLA is attractive because it is cheap and repeatable, but academic analysis, notably work presented at ASIACRYPT 2019, has shown that its standardized instantiation can both miss real leakage and flag leakage that no practical attack exploits. Attack-based evaluation, used in the Common Criteria smart-card domain and by the EMVCo security process, takes the opposite approach: the evaluator must actually recover the key and then report how much effort it took. Laboratories serving both regimes maintain equipment for both.

Timing analysis needs less hardware but more care. Cycle-accurate counters, logic analyzers, or time-to-digital converters measure execution time across many inputs, and statistical tests look for correlation between duration and secret-dependent branches or table lookups. Certification of a constant-time implementation is a claim that no such correlation exists across the whole input space, which is why timing measurements are repeated under varied temperature, supply voltage, and clock frequency. Thermal chambers and programmable supplies are consequently part of the side-channel bench, not just the environmental laboratory: a countermeasure that holds at 25 °C and nominal voltage may leak at the corners of the operating envelope.

Tamper and Fault Injection Testing

Physical security certification asks two questions: how hard is it to get inside, and what happens when someone tries. Tamper testing equipment answers both, evaluating passive resistance that slows an attacker down and active detection that destroys keys before the attacker reaches them.

Mechanical evaluation exercises drilling, milling, cutting, prying, and heat or solvent attacks against enclosures, potting compounds, and tamper meshes, generally under a stereo microscope so that the evaluator can work within a fraction of a millimeter of a sensing layer. High-speed instrumentation records the interval between mesh breach and completed zeroization, because a response that takes longer than the attacker needs to read out a key is no response at all. Battery-backed detection circuits are tested with the main supply removed and at low temperature, where the backup cell's capacity falls.

Semi-invasive and invasive analysis requires a semiconductor laboratory. Chemical decapsulation in a fume hood removes packaging; delayering and polishing expose metal layers for imaging in a scanning electron microscope; a focused ion beam mills through passivation to cut a trace or deposit a probe pad. Microprobing stations with low-capacitance probes and sub-micrometer positioning attempt to read bus traffic directly, and the evaluation records whether the active shield detected the intrusion. Fault injection is the complementary discipline: voltage and clock glitching systems drive the device outside its timing margins, electromagnetic pulse injectors couple transient fields into specific die regions, and near-infrared lasers inject faults through thinned backside silicon with micrometer spatial and nanosecond temporal precision. The objective is a controlled fault that skips an authentication check or corrupts a signature computation in an exploitable way.

Results are normalized through attack-potential scoring. The Joint Interpretation Library methodology used for smart cards and secure devices rates an attack on elapsed time, required expertise, knowledge of the target, window of opportunity, and equipment, separately for the identification phase and the exploitation phase. The summed score maps to the attack potential the product resists, and thence to the AVA_VAN level it can claim. This scoring is what makes physical evaluation comparable across laboratories despite the highly individual nature of each attack, and it is also what forces laboratories to keep expensive equipment current: an attack that required a focused ion beam a decade ago may now be within reach of a well-equipped hobbyist, and the scoring tables are revised accordingly.

Environmental Testing Systems

Security certifications require evidence that protective mechanisms work across the product's whole declared operating range, not only on the bench. Environmental test equipment supplies controlled, repeatable stress while the device under test continues to perform security functions and report its state.

Thermal chambers cover the commercial, industrial, and where relevant military ranges, with -40 °C to +85 °C typical for industrial equipment and wider ranges specified for defense and automotive products. Thermal cycling exposes marginal timing, solder fatigue, and temperature-dependent analog behavior. Security-specific testing inside the chamber verifies that cryptographic operations still produce correct results, that entropy sources retain their assessed min-entropy at the temperature extremes, and that tamper detection neither fails silently nor produces false trips. For FIPS 140-3 Level 4 modules the chamber is used to demonstrate environmental failure protection: the module must detect an excursion and zeroize, or be shown by testing to operate correctly throughout.

Humidity chambers evaluate corrosion, condensation, and surface leakage, usually in combined temperature-humidity profiles drawn from the IEC 60068 series for commercial equipment or MIL-STD-810 for defense equipment. Vibration tables and shock machines apply profiles representing transport and service conditions, and the security question is whether mechanical stress can open a tamper boundary, disturb a mesh connection, or provoke a false zeroization that renders the product unusable in the field.

Electromagnetic testing serves both compliance and security. Emissions testing in a semi-anechoic chamber or GTEM cell confirms that radiated and conducted emissions stay within regulatory limits, which also bounds the unintentional information leakage discussed in the side-channel section. Immunity testing applies electrostatic discharge per IEC 61000-4-2, radiated fields, and fast transients, and the security evaluation checks that interference cannot force the device into an insecure state, bypass an authentication step, or corrupt a key in memory without detection.

Operational Testing Platforms

Certification looks beyond the existence of a security feature to whether it holds up in use. Operational testing platforms recreate deployment conditions, workloads, and failure modes, and they capture what the security functions actually do under them.

Load testing subjects the product to realistic transaction volumes, concurrent sessions, and sustained operation. The security questions are specific: does the authentication path fall back to a weaker mechanism under load, does audit logging drop records when the buffer fills, does a rate limiter stop enforcing when the queue saturates. Automated harnesses generate representative traffic while instrumentation records authentication outcomes, access control decisions, and log completeness.

Interoperability testing verifies secure operation against diverse peers, versions, and configurations. It covers protocol version negotiation, cryptographic algorithm selection, and certificate path handling, and it deliberately includes downgrade scenarios, because an implementation that silently accepts an obsolete cipher suite fails a modern evaluation regardless of what it supports at its best. Reference implementations and competitor products are common test peers.

Lifecycle testing follows the product through delivery, installation, configuration, update, and decommissioning. Evaluators verify that keys are provisioned securely and never leave the boundary in plaintext, that the secure update mechanism authenticates firmware and refuses rollback to vulnerable versions, and that the guidance documentation is sufficient for a competent administrator to reach the evaluated configuration. Recovery testing uses programmable power controllers to interrupt operation at arbitrary points and then checks integrity verification at startup, the consistency of security state, and the behavior of the device after an interrupted update.

Documentation and Evidence Review Tools

High-assurance certification is as much an exercise in evidence management as in testing. A Common Criteria evaluation at a high AVA_VAN level or a FIPS 140-3 Level 3 submission can involve thousands of pages of design description, security policy, test procedure, and result, all of which must be internally consistent and traceable.

Requirements management systems maintain traceability matrices linking each requirement of the standard to the design element that satisfies it, the implementation component, the test case, and the recorded result. Automated analysis flags requirements with no test coverage, test cases with no requirement, and evidence that has gone stale relative to a changed specification. Evaluators depend on these matrices to demonstrate that their examination was complete rather than selective.

Version control and structured document comparison track evidence across the months or years an evaluation can span, so that an evaluator can confirm that a reported deficiency was corrected and that the correction did not disturb something already examined. Architecture analysis tools extract components, interfaces, information flows, and trust boundaries from design documentation and check them against the stated security policy. At the highest assurance levels formal methods tools verify that a security model satisfies the required properties and that a refinement of that model preserves them.

Audit and Compliance Tools

Most certifications are maintained rather than merely obtained. Configuration drift, an expired dependency, or an unpatched vulnerability can invalidate the evaluated configuration between formal assessments, so continuous monitoring tools fill the interval.

Configuration assessment tools compare running systems against an approved baseline and report unauthorized changes, missing patches, and insecure settings. Where the certification defines an evaluated configuration, these tools are what demonstrate that deployed instances remain inside it. Vulnerability management platforms correlate discovered weaknesses with asset inventories and compensating controls, and record remediation timelines and risk acceptance decisions, both of which an assessor will ask to see.

Log analysis and audit trail systems collect security-relevant events from distributed components, detect policy violations, and preserve the record. Tamper-evident storage and cryptographic signing of log entries protect the integrity of that record, which is the point: an audit trail that could have been edited proves nothing. Compliance reporting tools assemble test results, configuration snapshots, scan reports, and remediation evidence into the formats each scheme expects, and evidence repositories increasingly map a single control to the requirements of several schemes at once, since a product may hold Common Criteria, FIPS 140-3, and PCI approvals simultaneously with different renewal cycles.

Certification Laboratory Infrastructure

The laboratory itself is part of the evaluated system. Accreditation to ISO/IEC 17025 establishes technical competence, impartiality, and a functioning quality management system; scheme-specific licensing then adds the requirements of each program, such as NVLAP accreditation for CMVP testing, national scheme licensing for Common Criteria facilities, EMVCo qualification for type approval benches, and PCI recognition for PTS evaluations.

Physical facilities include environmentally controlled rooms with stable temperature, humidity, and power quality; uninterruptible supplies so that a multi-day test is not lost to an outage; and monitoring that documents conditions throughout an evaluation. Access control, surveillance, and safes protect evaluation targets, pre-release firmware, and source code, much of which is more sensitive than the shipping product.

Calibration is a continuous obligation. Every instrument whose reading contributes to a decision carries a calibration certificate traceable to national standards, with a stated uncertainty and an interval after which it must be recalibrated. Calibration management systems track those intervals and lock out equipment that has lapsed. Isolated test networks and segregated storage keep concurrent projects apart, both to prevent technical contamination and to honor the confidentiality obligations a laboratory owes to competing clients.

Personnel qualification completes the picture. Evaluators are trained in the scheme's methodology, in the specific instrumentation, and in the attack techniques they are expected to attempt, and several schemes require documented supervised experience before an evaluator can sign off independently. Because attack techniques improve continuously, competence maintenance is a standing requirement rather than an initial hurdle.

Emerging Certification Requirements

Certification requirements track the threat landscape, and three shifts are currently reshaping laboratory equipment.

Post-quantum cryptography is the most immediate. NIST published FIPS 203, 204, and 205 in August 2024, standardizing ML-KEM for key encapsulation, ML-DSA for lattice-based signatures, and SLH-DSA for hash-based signatures; a Falcon-based signature standard remains in draft, and HQC was selected in March 2025 as a code-based backup key encapsulation mechanism. Algorithm validation vectors for the new standards are already distributed through ACVP, so functional validation is largely a matter of updating harnesses. Side-channel evaluation is harder. Lattice schemes leak through operations with no classical analogue, including the number-theoretic transform, polynomial sampling, and the rejection sampling loop in ML-DSA whose iteration count itself depends on secret data. Leakage models, trace lengths, and countermeasure verification all had to be reworked, and laboratories evaluating post-quantum implementations need substantially deeper trace memory than an AES bench requires.

Connected-device certification is the second shift. Resource-constrained products cannot bear the cost of a full Common Criteria evaluation, so lighter-weight schemes have emerged: ETSI EN 303 645 provides a baseline for consumer IoT security, and GlobalPlatform's SESIP methodology together with the PSA Certified program offers a tiered, composable evaluation for platform components, so that a certified secure element can carry assurance upward into the product built on it. Testing emphasizes secure boot, firmware update authenticity, credential provisioning, and the wireless attack surface, all under battery-powered and intermittently connected conditions that complicate measurement.

Automotive cybersecurity is the third. ISO/SAE 21434 defines the engineering process, while UN Regulations 155 and 156 make a certified cybersecurity management system and software update management system a condition of vehicle type approval in the contracting parties to the relevant UNECE agreement. Laboratory equipment for this domain includes CAN, CAN FD, and automotive Ethernet interfaces for bus injection and fuzzing, electronic control unit test benches, hardware-in-the-loop rigs that let an attack be evaluated against a running vehicle model, and V2X radio test setups. Because a cybersecurity fault can become a safety fault, this testing has to be coordinated with functional safety analysis under ISO 26262 rather than performed in isolation.

Machine learning security is developing more slowly, and the certification frameworks are still forming. Current evaluation work focuses on adversarial robustness, model extraction, membership inference, and the integrity of training data and model weights, alongside conventional implementation security for the accelerators that run the models. Laboratories should expect requirements in this area to firm up over the next several years rather than treat present practice as settled.

Conclusion

Security certification equipment is the infrastructure that turns a security claim into a defensible finding. The specific instruments differ sharply by scheme, from the calibrated reference benches of EMV Level 1 approval to the focused ion beams of a smart-card vulnerability analysis, but the underlying discipline is the same: measure precisely, document completely, and make the result reproducible by someone else. As certification extends into post-quantum algorithms, connected devices, and vehicles, the equipment continues to grow more specialized, and the gap between owning a security feature and proving it works continues to widen.

Related Topics