Use the left and right arrow keys to move between lifecycle steps. While focus is inside the explorer, use Home to return to the complete diagram and End to reach the last step.
The complete lifecycle remains visible while the current step is highlighted.
Secure medical-device software development, step by step
Follow the complete CyberMed lifecycle from user needs through design validation. Each step connects IEC 62304 software processes with ANSI/AAMI SW96 security risk management and FDA's 2026 cybersecurity guidance.
How to use this explorer
Scroll to focus each box in the original diagram. The full lifecycle stays visible so you can follow feedback loops, verification paths, and risk decisions. Every step has a direct URL fragment.
System requirements translate user needs into measurable requirements for the complete medical device system.
What it is
The controlled set of functional, performance, safety, security, interface, environmental, and regulatory requirements for the full system.
Why it exists
Software requirements need a stable system context. Cybersecurity also depends on hardware, networks, cloud services, users, and related systems.
Main inputs
Approved user needs
System concept and architecture
Risk analyses
Regulatory and standards inputs
External interface constraints
Key activities
Define measurable system behavior
Allocate functions and controls to system elements
Specify interfaces and operating constraints
Establish verification methods and traceability
Expected outputs and evidence
System requirements specification
Allocation records
Interface requirements
Requirements traceability
Participating roles
Systems engineer, Product owner, Safety risk manager, Cybersecurity lead, Software and hardware leads
Lifecycle connection
System requirements come from user needs and allocate the software contribution that becomes software requirements. Design verification later checks the system design against them.
Software requirements define what the software must do, including the controls needed for safety and cybersecurity risk.
What it is
A verifiable specification for software functions, interfaces, data, performance, alarms, security, configuration, and risk controls.
Why it exists
Requirements create the traceable basis for architecture, implementation, verification, and change analysis.
Main inputs
Allocated system requirements
Safety and security risk controls
System architecture and interfaces
SOUP and platform constraints
Applicable regulatory inputs
Key activities
Define functional and nonfunctional software behavior
Include software risk-control requirements
Specify interfaces, data, and security objectives
Review requirements for clarity, consistency, testability, and traceability
Expected outputs and evidence
Software requirements specification
Security requirements
Traceability matrix
Requirements review record
Participating roles
Software lead, Systems engineer, Cybersecurity lead, Safety risk manager, Verification lead
Lifecycle connection
Software requirements are allocated from the system and feed software architecture. Risk loops return here when new or revised controls must become requirements.
Software architecture turns software requirements into a documented structure of software items and interfaces.
What it is
The top-level software structure, key software items, interfaces, responsibilities, dependencies, and segregation used to realize the requirements.
Why it exists
Architecture makes the system understandable enough to control complexity, isolate risk, plan integration, and verify design decisions.
Main inputs
Software requirements
System architecture
Safety and security risk controls
SOUP constraints
Platform and interface information
Key activities
Decompose software into items and interfaces
Define data flows and communication protocols
Specify SOUP needs and platform dependencies
Establish segregation needed for risk control
Verify the architecture against its inputs
Expected outputs and evidence
Software architecture document
Interface definitions
Segregation rationale
Architecture verification record
Participating roles
Software architect, Systems engineer, Cybersecurity architect, Software lead, Verification lead
Lifecycle connection
Software architecture follows software requirements and feeds both detailed software design and the security-architecture analysis shown in the left branch.
Threat modeling uses the architecture to identify assets, attack paths, threats, vulnerabilities, and candidate controls.
What it is
A repeatable analysis of how a capable threat actor could compromise the device, its related systems, or the clinical environment.
Why it exists
Threat modeling finds design weaknesses early and turns abstract security objectives into concrete risks and requirements.
Main inputs
Security architecture views
Intended use and foreseeable misuse
Asset and interface inventory
Threat intelligence
Known platform and supply-chain weaknesses
Key activities
Define scope and system boundaries
Identify assets, threat actors, and attack surfaces
Analyze abuse cases and attack paths
Record threats and potential vulnerabilities
Propose controls and trace findings into risk management
Expected outputs and evidence
Threat-model report
Data-flow diagrams
Threat and attack-path register
Assumptions and scope record
Trace links to risks and controls
Participating roles
Threat-model facilitator, Cybersecurity engineer, Software and systems architects, Clinical or domain expert, Risk manager
Lifecycle connection
Threat modeling consumes the security views and produces risk scenarios for estimation and evaluation. It is updated when design or threat information changes.
The assessment turns threat-model findings into an acceptability decision. Unacceptable risks move to control selection; acceptable risks remain documented residual risks.
A yes path permits detailed software design to proceed with documented residual risk. A no path leads to cybersecurity controls and revised requirements or architecture.
Software design follows architecture and provides the blueprint for implementation. Risk loops may return here when testing or analysis exposes a design weakness.
An acceptable result permits unit verification to continue. An unacceptable result returns to design, component selection, or implementation before the build advances.
Unit verification checks each software unit against defined acceptance criteria before or during integration.
What it is
Risk-based review, analysis, and testing of individual software units, including their expected behavior, interfaces, error handling, and coding quality.
Why it exists
Defects are less costly to diagnose at unit level, and critical control logic needs direct evidence before higher-level tests obscure the cause.
Main inputs
Implemented software units
Detailed design
Unit acceptance criteria
Coding standards
Risk-control requirements
Key activities
Define and review unit-verification methods
Test normal, boundary, and error behavior
Review code against acceptance criteria
Verify unit-level security controls
Record and resolve anomalies
Expected outputs and evidence
Unit test procedures and results
Coverage or analysis records where justified
Code-review results
Anomaly and resolution records
Traceability to design and controls
Participating roles
Software developers, Unit-test engineer, Software lead, Cybersecurity reviewer, Independent reviewer where required by procedure
Lifecycle connection
Unit testing verifies implementation against detailed design. Results support integration and are complemented by static and dynamic security analysis.
Acceptable findings allow integration testing to continue. Unacceptable findings return to software design or implementation for correction and re-verification.
Integration testing follows unit-level verification and security analysis. Its evidence supports software verification against the full requirements set.
Software verification closes the software side of the V-model against software requirements. Its residual anomalies then receive a focused security assessment.
The assessment follows software verification. Its findings lead to an acceptability decision before cybersecurity testing and higher-level verification continue.
The release team decides whether the security risk from each unresolved anomaly is acceptable under the approved criteria.
What it is
A release-facing risk decision that either permits a documented anomaly to remain or sends corrective work back into requirements, design, implementation, or verification.
Why it exists
A controlled decision prevents the schedule from silently overriding security risk management.
Main inputs
Anomaly security assessment
Approved risk criteria
Software verification evidence
Potential safety impact
Proposed monitoring or remediation
Key activities
Evaluate each security-relevant anomaly
Confirm residual-risk rationale
Review cumulative and interacting effects
Approve corrective action where risk is unacceptable
An acceptable result opens the cybersecurity-testing branch. An unacceptable result returns to software requirements or design for corrective work and repeat verification.
Cybersecurity testing challenges the integrated device system, its controls, and its exposed attack surface.
What it is
Risk-based testing that combines control verification, vulnerability testing, misuse and abuse cases, fuzzing, penetration testing, and other suitable methods.
Why it exists
Security architecture and implementation claims need evidence under realistic and adversarial conditions, not only normal functional use.
Main inputs
Release-candidate device system
Threat model and attack surface
Security requirements and controls
Architecture views
Known vulnerabilities and abuse cases
Key activities
Define risk-based scope and independence
Verify control implementation and effectiveness
Exercise misuse, abuse, and malformed-input cases
Perform vulnerability and penetration testing
Document methods, configurations, results, and limitations
Expected outputs and evidence
Cybersecurity test plan and report
Control-verification results
Penetration or vulnerability-test report
Test environment and tool records
Finding traceability
Participating roles
Security test lead, Penetration tester, Product security engineer, Software and systems engineers, Security risk manager
Lifecycle connection
Cybersecurity testing follows acceptable anomaly risk and challenges the integrated system. Discovered vulnerabilities feed another risk loop before design verification.
Discovery follows cybersecurity testing and supplies the final risk-estimation loop. Confirmed weaknesses may also update the threat model and test coverage.
The team evaluates vulnerabilities found during system-level cybersecurity testing in the device and clinical context.
What it is
A focused security risk assessment of confirmed test findings, including exploitability, adverse impact, safety impact, affected versions, and control effectiveness.
Why it exists
Test severity labels alone do not determine medical-device risk. The device architecture, use environment, and possible harm shape the decision.
Main inputs
Controlled vulnerability records
Test and reproduction evidence
Threat model and architecture
Risk criteria
Safety and clinical impact analysis
Key activities
Map each vulnerability to assets and threat scenarios
A yes path advances to design verification. A no path returns to software design and implementation, followed by all affected verification and security tests.
Design verification confirms that system design outputs meet the approved system requirements.
What it is
The system-level reviews, analyses, and tests that show the implemented device conforms to its design inputs, including cybersecurity requirements.
Why it exists
System verification checks behavior across software, hardware, services, interfaces, and controls that software verification cannot cover alone.
Main inputs
System requirements
Integrated device system
System verification plan
Software and cybersecurity evidence
Risk-control traceability
Key activities
Execute system-level verification methods
Verify interfaces and related-system behavior
Confirm system security requirements and controls
Evaluate results and anomalies
Complete traceability and change impact
Expected outputs and evidence
Design verification report
System test and analysis records
Requirements coverage
System risk-control verification
Anomaly disposition
Participating roles
Systems verification lead, Systems engineer, Cybersecurity test lead, Safety risk manager, Quality representative
Lifecycle connection
Design verification follows acceptable software and cybersecurity evidence and closes the loop to system requirements. It supplies a verified device for design validation.
Design validation confirms that the finished device meets user needs and intended-use requirements in representative conditions.
What it is
The planned evaluation of the complete device with production-equivalent units, representative users, intended environments, and applicable clinical or use scenarios.
Why it exists
A device can meet written requirements and still fail to support safe, effective, and secure use in the real environment.
Main inputs
Approved user needs and intended use
Verified device system
Validation plan and acceptance criteria
Representative users and environments
Residual-risk and labeling information
Key activities
Validate intended use and key user needs
Use representative conditions and production-equivalent units
Confirm secure configuration and operational assumptions
Evaluate unexpected results and use-related effects
Complete final traceability and approvals
Expected outputs and evidence
Design validation protocol and report
User-need coverage
Representative-environment evidence
Residual-risk and labeling confirmation
Final approval records
Participating roles
Validation lead, Clinical and human factors specialists, Systems engineer, Product security representative, Quality and regulatory representatives
Lifecycle connection
Design validation closes the top V-model loop to user needs. Its findings may restart any affected lifecycle activity through controlled change and risk management.
Every step cites the clause or section it relies on. Each citation has a label that tells you how much weight it carries. The diagram and the step order are CyberMed's integration of these sources. No single standard defines this lifecycle.
Published . Updated . Written and reviewed under the CyberMed editorial policy.
Source labels
Statutory requirement
A requirement in the Federal Food, Drug, and Cosmetic Act. It is law for the submissions and devices that the statute covers.
Normative requirement
A requirement stated in the cited clause of a consensus standard. It applies when a manufacturer claims conformance to that standard. Some IEC 62304 clauses apply only to software safety Class B or Class C.
FDA guidance
A recommendation in FDA's premarket cybersecurity guidance. Guidance describes FDA's current thinking and is not binding, except where it restates a statutory or regulatory requirement.
Recommended practice
Informative material in a standard, or a common engineering practice. It is useful, but no cited source requires it.
CyberMed explanation
CyberMed's reading of how the step fits the lifecycle diagram. It is an interpretation, not a requirement from any source.