CSSLP Exam Prep · Topic 1 of 5

Secure Software Concepts & Requirements

CIA Triad · Security Principles · Compliance · Privacy by Design · Data Classification

Study with Practice Tests →

Secure Software Concepts & Requirements

Domains 1 & 2 of the CSSLP CBK (~24% combined). These foundational domains establish the security mindset every secure software professional must apply throughout the entire software lifecycle.

~24% of Exam · Domains 1 & 2

CSSLP Exam Overview — All 8 Domains

DomainWeightKey Topics
1. Secure Software Concepts10%CIA triad, security principles, risk, trust models
2. Secure Software Requirements14%Requirements gathering, privacy, compliance, abuse cases
3. Secure Software Design14%Threat modeling, STRIDE, security architecture
4. Secure Software Implementation14%Secure coding, common vulns, code review
5. Secure Software Testing14%SAST, DAST, IAST, fuzzing, pen testing
6. Secure Software Lifecycle Management11%SDLC integration, DevSecOps, metrics
7. Secure Software Deployment & Ops11%Deployment, configuration, patch management, IR
8. Secure Software Supply Chain12%Third-party risk, SBOM, CI/CD security

Core Concepts at a Glance

CIA Triad

Confidentiality: Only authorized parties can access data (encryption, access controls).
Integrity: Data is accurate and unmodified (hashing, digital signatures).
Availability: Systems accessible when needed (redundancy, DDoS protection).

Security Design Principles

Least Privilege: Minimum access needed.
Defense in Depth: Multiple overlapping layers.
Fail Secure: Deny on failure.
Economy of Mechanism: Keep it simple.
Separation of Duties: No single point of control.

Security Requirements

Security requirements must be defined before design begins. Sources: business objectives, legal/regulatory mandates, use cases + abuse cases, stakeholder interviews, security standards (OWASP ASVS, ISO 27001), and threat modeling outputs.

Privacy by Design

7 Foundational Principles: Proactive not Reactive · Privacy as Default · Embedded into Design · Full Functionality · End-to-End Security · Visibility/Transparency · Respect for User Privacy. Privacy must be built in — not bolted on.

Data Classification

Determines security controls applied. Typical levels: Public → Internal → Confidential → Restricted/Top Secret. Classification drives: encryption requirements, access control stringency, retention policies, and disposal methods.

Abuse Cases

The inverse of use cases. Describe how a malicious actor could misuse the system. Used to derive security requirements from attacker perspective. More effective than simple use cases for identifying security gaps early in the SDLC.

Exam Tips

Principle Questions

The CSSLP loves asking which principle applies to a given scenario. "Economy of mechanism" = keep security designs simple. "Complete mediation" = check every access, every time. "Psychological acceptability" = security mechanisms shouldn't make the system harder to use than without them.

Requirements Timing

Security requirements must be gathered at the start of the SDLC. Adding security after design = bolt-on security = expensive and less effective. The exam will try to trick you into thinking security can be retrofitted — it cannot effectively be.

CIA vs AAA

CIA = data security properties (Confidentiality, Integrity, Availability). AAA = access control framework (Authentication, Authorization, Accounting/Auditing). Both are foundational — know which applies to which scenario.

Core Security Concepts

Domain 1 (~10%) establishes the foundational security vocabulary and principles that underpin every other CSSLP domain.

The CIA Triad — Extended

PropertyDefinitionThreatsControls
ConfidentialityPreventing unauthorized disclosure of informationEavesdropping, data breach, shoulder surfingEncryption, access controls, data masking
IntegrityEnsuring data accuracy and preventing unauthorized modificationTampering, man-in-the-middle, SQL injectionHashing (SHA-256), digital signatures, input validation
AvailabilityEnsuring systems and data are accessible when neededDDoS, ransomware, hardware failureRedundancy, load balancing, backups, DDoS mitigation

AAA Framework

Authentication

Proving identity. Factors: Something you know (password), Something you have (token/OTP), Something you are (biometric), Somewhere you are (location). MFA combines two or more.

Authorization

Determining what an authenticated identity is permitted to do. Models: RBAC (Role-Based), ABAC (Attribute-Based), DAC (Discretionary), MAC (Mandatory). Principle: deny by default, grant explicitly.

Accounting (Auditing)

Recording what authenticated and authorized users actually did. Enables forensics, compliance, anomaly detection, and non-repudiation. Logs must be tamper-evident (append-only, integrity protected).

Foundational Security Design Principles

PrincipleDefinitionExample Application
Least PrivilegeGrant minimum access rights needed for a functionDatabase user has SELECT only, not DROP TABLE
Separation of DutiesNo single person/process controls a critical function end-to-endDeveloper cannot deploy to production
Defense in DepthMultiple layered security controls; no single point of failureWAF + input validation + parameterized queries
Fail SecureOn failure, system defaults to a secure (denying) stateFirewall drops all traffic if rules engine crashes
Economy of MechanismKeep security designs as simple as possiblePrefer simple crypto libraries over custom implementations
Complete MediationCheck every access request every time; no caching of decisionsRe-validate session on every sensitive operation
Open DesignSecurity should not depend on secrecy of design (Kerckhoffs)Security of AES relies on key secrecy, not algorithm secrecy
Psychological AcceptabilitySecurity mechanisms must not make system harder to useSSO reduces password fatigue while maintaining security
Least Common MechanismMinimize shared mechanisms between usersSeparate database connections per tenant

Trust Models

Zero Trust

"Never trust, always verify." No implicit trust based on network location. Every request authenticated and authorized regardless of source. Key tenets: verify identity explicitly, use least privilege access, assume breach.

Implicit Trust (Traditional)

Once inside the network perimeter, entities are trusted. Dangerous in modern environments with remote work, cloud, and insider threats. The basis of the "castle-and-moat" security model now considered inadequate.

Risk & Risk Management

Risk = Likelihood × Impact. Risk responses: Accept, Avoid, Transfer (insurance), Mitigate (controls). Residual risk = risk remaining after controls. Risk appetite = amount of risk an organization is willing to accept.

Secure Software Requirements

Domain 2 (~14%). Security requirements define what the software must do (and must not allow) from a security perspective. They must be defined before design begins.

Types of Security Requirements

Functional Security Requirements

Specific security behaviors the system must implement. Examples: "The system shall enforce MFA for all admin accounts," "The system shall encrypt all PII at rest using AES-256," "Passwords shall be hashed using bcrypt with a cost factor ≥ 12."

Non-Functional Security Requirements

Security quality attributes not tied to specific functions. Examples: "The authentication system shall respond within 500ms," "The system shall achieve 99.99% availability," "All API endpoints shall support TLS 1.2 minimum."

Derived Requirements

Requirements derived from threat modeling, compliance mandates, or system architecture decisions. Not explicitly stated by stakeholders but necessary to meet stated goals. Often discovered during threat modeling (STRIDE analysis).

Requirements Elicitation Sources

SourceExamplesTechniques
StakeholdersBusiness owners, legal, compliance, end usersInterviews, workshops, surveys
Regulatory/LegalGDPR, HIPAA, PCI-DSS, SOX, CCPACompliance gap analysis
StandardsOWASP ASVS, NIST SP 800-53, ISO 27001Requirements mapping
Threat ModelingSTRIDE analysis outputsAttack tree analysis
Abuse CasesAttacker scenarios, misuse casesNegative use case analysis
Prior IncidentsVulnerability history, breach post-mortemsLessons-learned review

Use Cases vs Abuse Cases

Use Case (Positive)

"A registered user can log in with their email and password."
Describes intended, legitimate interactions with the system. Forms the basis of functional requirements. Security teams extend these into abuse cases.

Abuse Case (Negative)

"An attacker attempts to log in by brute-forcing passwords."
Describes how a malicious actor misuses the same functionality. Directly generates security requirements: rate limiting, account lockout, CAPTCHA, MFA.

Misuse Case

Broader than an abuse case — includes accidental misuse by legitimate users, not just malicious intent. Example: "A user accidentally exports an entire customer database instead of their own records." Drives access scoping requirements.

Data Classification

LevelDescriptionExamplesControls Required
PublicIntentionally shared with anyoneMarketing materials, press releasesIntegrity controls only
InternalFor internal use onlyEmployee handbook, meeting notesAccess controls, limited sharing
ConfidentialSensitive business dataFinancial data, client lists, IPEncryption at rest/transit, strict access
RestrictedHighest sensitivity; regulatory or legal protectionPII, PHI, PCI data, trade secretsStrong encryption, audit logging, MFA, DLP

OWASP Application Security Verification Standard (ASVS)

Level 1 — Opportunistic

Basic security. Verifiable by black-box testing. Minimum bar for all software. Covers most common OWASP Top 10 vulnerabilities. Appropriate for: low-risk applications, first pass on all software.

Level 2 — Standard

Standard for most applications. Requires security controls that defend against the majority of risks today. Appropriate for: applications handling sensitive data, business-critical systems.

Level 3 — Advanced

Highest assurance level. Requires full documentation and source access. Appropriate for: critical infrastructure, military, medical devices, financial systems where failure could cause significant harm.

Compliance & Privacy

Regulatory compliance and privacy requirements are major drivers of software security requirements. Understanding key frameworks helps you map legal obligations to technical controls.

Key Compliance Frameworks

FrameworkScopeKey RequirementsPenalty for Non-Compliance
GDPREU citizens' personal data, globallyConsent, right to erasure, data minimization, breach notification (72 hrs), DPO, Privacy by DesignUp to 4% global annual revenue or €20M
HIPAAUS healthcare — PHI (Protected Health Information)Administrative, physical, technical safeguards; breach notification; Business Associate Agreements$100–$50,000 per violation
PCI-DSSPayment card data (anyone storing/processing/transmitting)12 requirements: network security, cardholder data protection, vulnerability management, access control, monitoringFines, loss of card processing rights
SOXUS publicly traded companies — financial reportingSections 302 (CEO/CFO attestation), 404 (internal controls), audit trails, access controls for financial systemsCriminal penalties, fines, delisting
CCPACalifornia consumers' personal dataRight to know, opt-out of sale, non-discrimination, deletion rights$2,500–$7,500 per intentional violation
FISMAUS federal agencies and contractorsRisk management framework (NIST RMF), continuous monitoring, ATO (Authority to Operate)Loss of federal contracts

Privacy by Design — 7 Principles

1. Proactive not Reactive

Anticipate and prevent privacy risks before they occur. Privacy embedded from the start — not discovered after the fact during audits or incidents.

2. Privacy as the Default Setting

Maximum privacy protection is the default — users shouldn't have to take action to protect their privacy. Data minimization by default.

3. Privacy Embedded into Design

Privacy is a core feature, not an add-on. It is integrated into the system architecture and business practices seamlessly.

4. Full Functionality

Privacy and security achieve all legitimate objectives — no unnecessary trade-offs. Avoid false "privacy vs. usability" dichotomies.

5. End-to-End Security

Strong security measures throughout the entire lifecycle — from collection to retention to deletion. Secure destruction when data is no longer needed.

6. Visibility & Transparency

Components and operations remain visible and transparent — open to independent verification. No hidden agendas or secret data practices.

7. Respect for User Privacy

Keep the system user-centric. Provide strong privacy defaults, appropriate notice, and genuine user empowerment to control their own data.

Privacy Concepts in Software

ConceptDefinitionImplementation
Data MinimizationCollect only data necessary for the stated purposeRemove optional fields; regularly audit what data is stored
Purpose LimitationUse data only for the purpose it was collectedSeparate data stores per purpose; access controls by purpose
AnonymizationRemove all identifying information irreversiblyGeneralization, suppression, noise addition
PseudonymizationReplace identifying info with pseudonyms (reversible with key)Tokenization, key-based substitution; still considered personal data
Consent ManagementUsers explicitly agree to data collection and processingGranular opt-in checkboxes, consent database, withdrawal mechanism
Right to Erasure"Right to be forgotten" — users can request deletionData deletion workflows, cascade deletes, backup purging processes

Practice Quiz — Secure Software Concepts & Requirements

10 questions covering CIA triad, security principles, compliance frameworks, privacy, and security requirements. Select the best answer.

1. Which security principle states that users and processes should be granted only the minimum access rights necessary to perform their function?
2. A Denial of Service attack primarily targets which component of the CIA triad?
3. The "fail secure" design principle means that when a system fails, it should:
4. Which compliance framework applies specifically to organizations that store, process, or transmit credit card data?
5. An abuse case differs from a use case in that it:
6. Which security design principle emphasizes keeping security mechanisms as simple as possible?
7. Privacy by Design requires that privacy controls be:
8. Which data classification level typically requires the strictest access controls, mandatory encryption, and audit logging?
9. The AAA security framework stands for:
10. Which regulation specifically protects the personal data of European Union citizens and mandates breach notification within 72 hours?

Memory Hooks

Six sticky mental anchors for the highest-yield Concepts & Requirements topics.

🔐
CIA Triad
"Confidentiality = Lock · Integrity = Hash · Availability = Uptime"
DoS attacks → Availability. Data breach → Confidentiality. Tampering → Integrity. On the exam, identify which CIA component is threatened and map controls accordingly. A hash protects integrity but not confidentiality.
🛡️
Security Design Principles
"Least · Separated · Defense · Fail Secure · Simple"
Least Privilege, Separation of Duties, Defense in Depth, Fail Secure, Economy of Mechanism (keep simple). For the exam: "Economy of Mechanism" = simplicity. "Complete Mediation" = check every access every time. "Open Design" = no security through obscurity.
⚖️
Risk Formula
Risk = Likelihood × Impact
Four responses: Accept (tolerate), Avoid (eliminate), Transfer (insurance), Mitigate (add controls). Residual risk = what's left after mitigation. Risk appetite = how much risk the org is willing to accept. Never "eliminate" risk completely.
🗂️
Abuse Cases
"Use Case = Intended · Abuse Case = Attacker's View"
For every use case "User logs in," write the abuse case "Attacker brute-forces login." The abuse case generates security requirements: rate limiting, lockout, MFA. Start from abuse cases early — fixing requirements is 100x cheaper than fixing deployed code.
🔏
Privacy by Design
"Proactive · Default Private · Embedded · Full Function · End-to-End · Transparent · User-Centric"
7 principles — but the CSSLP exam focuses on two: Privacy as the DEFAULT (max privacy without user action) and Privacy EMBEDDED in design (not bolted on). GDPR codifies Privacy by Design in Article 25: "data protection by design and by default."
📋
Compliance Anchors
"GDPR=EU · HIPAA=Health · PCI=Cards · SOX=Finance"
GDPR: EU personal data, 72-hour breach notification, right to erasure. HIPAA: US healthcare PHI, HITECH extends to business associates. PCI-DSS: payment card data, 12 requirements. SOX: financial reporting controls for public companies (Sections 302 and 404 are key).

Flashcards & Study Advisor

Click any card to flip it. 8 high-yield concept cards for rapid review.

👆 Click a card to reveal the answer

Complete Mediation
What does this principle require on every access?
Every access to every resource must be checked every time — no caching of access decisions. If a user's permissions change, the next request must reflect the new rights immediately. Prevents privilege escalation via cached permissions.
Pseudonymization vs Anonymization
Which is reversible? Which is still personal data under GDPR?
Pseudonymization: replaces identifiers with pseudonyms (reversible with key). Still considered personal data under GDPR — still regulated. Anonymization: irreversible removal of all identifying info. No longer personal data under GDPR — not regulated. Tokenization is a form of pseudonymization.
GDPR Key Requirements
72 hours, right to erasure, and what else?
Breach notification within 72 hours to supervisory authority. Right to erasure ("right to be forgotten"). Data minimization. Purpose limitation. Privacy by design (Art. 25). Data Protection Officer (DPO) for certain orgs. Lawful basis for processing. Data Subject Access Requests (DSAR).
OWASP ASVS Levels
What do Levels 1, 2, and 3 represent?
Level 1 — Opportunistic: basic security, black-box verifiable, minimum for all software.
Level 2 — Standard: most applications, defends majority of risks, requires grey-box testing.
Level 3 — Advanced: critical systems (medical, finance), requires full code access and documentation.
Kerckhoffs's Principle
What does "open design" mean for cryptography?
A cryptosystem should be secure even if everything about the system — except the key — is public knowledge. Security must rely on key secrecy, not algorithm secrecy. AES is public; its security comes from the key. "Security through obscurity" violates this principle.
Risk Response Types
What are the four ways to respond to identified risk?
Accept: tolerate the risk (residual risk within appetite).
Avoid: eliminate the risk by stopping the risky activity.
Transfer: shift risk to third party (insurance, contractual liability).
Mitigate: implement controls to reduce likelihood or impact.
Note: risk can never be fully eliminated — only reduced.
PCI-DSS 12 Requirements
What are the 6 major goals of PCI-DSS?
1. Build/maintain secure network (firewall, no vendor defaults).
2. Protect cardholder data (encryption at rest/transit).
3. Vulnerability management (patching, AV).
4. Strong access control (need-to-know, unique IDs, physical).
5. Monitor/test networks (logging, pen testing).
6. Information security policy.
Separation of Duties
Why is it a critical control and where does it apply in software?
No single individual should control an entire critical process end-to-end — prevents fraud and errors. In software: developers cannot approve their own code to production (4-eyes principle), write access to prod requires a change ticket and second approval. SOX Section 404 mandates this for financial systems.

Study Advisor

Select a topic for targeted exam-day guidance.

CIA Triad & AAA — Exam Strategy

  • Map every attack to a CIA component: ransomware = Availability + Confidentiality; SQLi = Confidentiality + Integrity; DDoS = Availability only.
  • CIA is about data properties; AAA is about access control. Know which framework the question is referring to.
  • Non-repudiation ties to Integrity and Accounting — digital signatures provide both integrity and non-repudiation.
  • Multi-factor authentication strengthens the Authentication "A" — combines factors (know/have/are).
  • Audit logs = Accounting "A" — must be tamper-evident; send to a separate log server the application cannot write to.

Ready to pass the CSSLP?

Practice with full-length adaptive exams covering all 8 domains.

Start Free Practice Tests →