CISSP Cheat Sheet — One-Week Revision

All 8 Domains · Domain-Wise Organisation · Comprehensive Quick Reference
📚 Official (ISC)² CBK Aligned 📄 ~110–120 Pages Printed · D1≈22 · D2≈12 · D3≈26 · D4≈13 · D5≈11 · D6≈11 · D7≈17 · D8≈11 🎯 Revision Aid (Not Teaching Material)

📅 Suggested 1-Week Revision Schedule

DayDomainsExam WeightFocus Areas
MondayD1 + D226%Risk, Governance, Compliance, Legal, BCP, Data Classification, Privacy
TuesdayD313%Security Models, Cryptography, PKI, System Architecture, Cloud
WednesdayD413%OSI/TCP-IP, Protocols, Firewalls, VPN, Wireless, Network Attacks
ThursdayD513%Authentication, Biometrics, SSO, Kerberos, SAML/OAuth, Authorization
FridayD6 + D822%Pen Testing, Audits, SOC Reports, SDLC, Secure Coding, OWASP
SaturdayD713%Incident Response, Forensics, SIEM, DRP, Backup, Physical Security
SundayCross-Review100%Formulas, Mnemonics, Exam Traps, Weak Areas, Brain Dump Practice

📑 Master Table of Contents

Domain 1: Security & Risk Management

16% of Exam · Heaviest Domain · Priority ★★★★★

1.1 Professional Ethics

(ISC)² Code of Ethics — Four Canons (Priority Order)

#CanonObligationKey Point
1Protect SocietyCommon good, public trust, infrastructureOverrides ALL other canons — report even if employer objects
2Act HonourablyHonestly, justly, responsibly, legallyTowards principals, clients, employers
3Provide Diligent ServiceCompetent service to principalsDon't practice outside your competence
4Advance the ProfessionPersonal development, professional integritySelf-interest is ALWAYS last
🧠 Mnemonic — "SAPA": Society → Act honourably → Provide service → Advance profession. "Society Always Precedes Ambition."

Ethical Frameworks

  • Computer Ethics Institute 10 Commandments: No harm, no interference, no snooping, no theft, no false testimony, pay for software, respect IP, think of social consequences
  • RFC 1087 — Unethical Internet Use: Unauthorised access, disruption, waste of resources, integrity violations, privacy violations
  • Responsible Disclosure: Notify vendor privately → allow time to patch → coordinate disclosure → publish if unresponsive (Google: 90 days)

Organizational Code of Ethics

Separate from (ISC)² Code: Each organisation should have its own code of ethics aligned to its mission. Security professionals must adhere to both the (ISC)² Code and their organisational code. Where they conflict, the (ISC)² Code takes precedence (Canon 1 — protect society).
⚠️ Exam Trap: If canons conflict, higher-numbered canon yields. Protecting society (Canon 1) overrides loyalty to employer (Canon 3). Whistleblowing is justified when public safety is at stake.

1.2 Security Concepts — CIA Triad & Beyond

🔑 5 Pillars of Information Security (ISC² 2024): Confidentiality · Integrity · Availability · Authenticity · Non-repudiation. The exam uses this exact "5 Pillars" label.

CIA Triad

PropertyGoalAttacks (DAD)Controls
ConfidentialityPrevent unauthorised disclosureDisclosureEncryption, access controls, classification
IntegrityPrevent unauthorised modificationAlterationHashing, digital signatures, change mgmt
AvailabilityEnsure timely accessDestruction/DoSRedundancy, backups, DR plans

Additional Concepts

ConceptMeaning
Non-repudiationCannot deny action (digital signatures)
AuthenticityVerified identity of sender/data
AccountabilityActions traceable to individual (audit logs)
Defense in DepthMultiple layered controls
🔑 CISSP Priority: Banks prioritise Integrity; streaming services prioritise Availability; law firms prioritise Confidentiality. The exam tests whether you can identify the primary CIA property at stake.
⚠️ Exam Trap: Ransomware attacks all three CIA properties, but the primary target is Availability (to extort payment). Read "primarily" carefully.

1.3 Security Governance Principles

Governance vs. Management vs. Operations

LevelWhoDoes WhatArtefacts
GovernanceBoard / Senior MgmtDirection, oversight, risk appetitePolicies, charters
ManagementCISO / DirectorsPlanning, resource allocation, controlsStandards, procedures
OperationsIT / Security TeamsExecution, monitoring, responseLogs, tickets, configs

Key Security Roles

RoleResponsibilityKey Distinction
Data OwnerClassifies data, approves access, sets policyBusiness executive — NOT IT
Data CustodianImplements controls, backups, maintenanceIT/sysadmin — follows Owner's decisions
Data StewardData quality, metadata, complianceDay-to-day governance
Data ProcessorProcesses data on behalf of controllerGDPR term — think cloud provider
Data ControllerDetermines purpose of processingGDPR term — think your organisation

Organisational Processes

ProcessSecurity Impact
AcquisitionsDue diligence on target's security posture; inherited risk/compliance debt; data integration classification
DivestituresData separation; access revocation; IP protection; contractual security obligations to buyer
Governance CommitteesBoard-level risk committee; security steering committee; cross-functional oversight of security strategy

Key Governance Frameworks

FrameworkFocusKey Feature
NIST CSF 2.0Cybersecurity risk6 Functions: Govern→Identify→Protect→Detect→Respond→Recover
ISO 27001ISMSCertifiable; 10 clauses + Annex A (93 controls)
ISO 27002Control guidanceNOT certifiable — guidance for Annex A
COBITIT governanceISACA; aligns IT with business goals
NIST RMFRisk management7 Steps: Prepare→Categorise→Select→Implement→Assess→Authorise→Monitor
NIST SP 800-53Security & privacy controlsComprehensive control catalog (20 families); used by RMF control selection step
NIST SP 800-171CUI protectionNon-federal orgs handling Controlled Unclassified Information; 110 controls
TOGAF / SABSAEA / Security architectureEnterprise architecture frameworks
FedRAMPCloud for US govtStandardised cloud security assessment; based on NIST 800-53; Authorise → Monitor
Due Care = doing the right thing (acting prudently). Due Diligence = verifying it works (researching, auditing). Care = Act, Diligence = Check.
⚠️ Exam Trap: Data Owner is a business executive, never IT staff. Senior management approves policies (not the CISO alone). ISO 27002 is NOT certifiable.

1.4 Compliance & Regulatory Requirements

RegulationScopeKey RequirementsPenalties / Notes
GDPREU personal data7 principles; 72-hr breach notification; DPO required for large-scale processingUp to 4% global revenue or €20M
HIPAAUS healthcare (PHI)Privacy Rule + Security Rule + Breach Notification60-day breach notification to HHS
SOXUS public companiesSec 302/404: internal controls over financial reportingCEO/CFO personal criminal liability
PCI-DSSPayment card data6 goals, 12 requirementsContractual (not law); fines via card brands
GLBAUS financial servicesPrivacy + Safeguards + Pretexting RulesFinancial institution security plans
FERPAUS student recordsParental consent for disclosureLoss of federal funding
FISMAUS federal agenciesNIST RMF mandatoryAnnual security reports
CCPA/CPRACalifornia consumersRight to Know/Delete/Opt-Out/Correct$7,500/intentional violation
PIPLChina personal informationConsent-based; data localisation; cross-border transfer restrictionsUp to 5% annual revenue; personal liability for DPOs
POPIASouth Africa personal info8 conditions for lawful processing; Information Regulator oversightFines up to ZAR 10M or imprisonment
🔑 Compliance is the floor, not the ceiling. Meeting regulatory requirements does NOT equal being secure. A compliant organisation can still be breached.

Breach Notification Timelines

RegulationNotifyTimeframe
GDPRSupervisory authority72 hours
HIPAAHHS (+ individuals if >500)60 days
PCI-DSSCard brands + acquirerASAP (contractual)

1.5 Legal & Regulatory Issues

Legal System Types

SystemBasisWhere
Common LawPrecedent + judicial interpretationUK, US, Australia
Civil/Code LawComprehensive written legal codesContinental Europe, Latin America
Religious LawFaith-based (Sharia, Canon)Some Middle Eastern / Vatican
Customary LawTradition-basedParts of Africa, Asia

Law Categories

TypeWho InitiatesBurden of ProofOutcome
CriminalGovernment (prosecutor)Beyond reasonable doubtImprisonment, fines
Civil (Tort)Victim (plaintiff)Preponderance of evidence (51%+)Money damages
AdministrativeRegulatory agencySubstantial evidenceLicence revocation, fines

Negligence — Four Elements (ALL required)

DutyBreachCausationDamages

All four must be proven. If any element is missing, no negligence claim.

Intellectual Property

TypeProtectsDurationKey Distinction
CopyrightCreative expression (code, books, art)Life + 70 years (individual);
Work-for-hire: 95 yrs from pub / 120 yrs from creation
Automatic; protects expression NOT ideas
TrademarkBrand identity (logos, names)Unlimited (if maintained)™ = claimed; ® = registered
PatentNovel inventions20 yrs (utility) / 15 yrs (design)Requires public disclosure
Trade SecretConfidential business infoUnlimited (if actively protected)No registration — lost if disclosed

Open Source Licence Risk

LicenceCopyleft?Commercial Risk
GPLStrong (viral)HIGH — derivative works must be GPL
LGPLWeak (library only)MEDIUM — library changes shared
MIT / Apache / BSDNone (permissive)LOW — free for commercial use
Work for Hire: IP created by employees during employment scope belongs to the employer. Contractors require explicit IP assignment agreements.

Evidence Admissibility

Evidence must be: Relevant + Reliable + Legally Obtained

Business Records Exception: System logs kept in the ordinary course of business are admissible despite the hearsay rule.

CFAA (Computer Fraud and Abuse Act): Criminalises unauthorised access to protected computers; both civil & criminal remedies.

⚠️ Exam Trap: Copyright protects expression not ideas. Trade secrets require active protection or they are lost forever. GPL "infects" proprietary code that links to it.

Import/Export Controls

RegulationScopeKey Point
EAR (Export Administration Regulations)Dual-use items (commercial + military potential)Administered by Bureau of Industry & Security (BIS); includes encryption software
ITAR (International Traffic in Arms Regulations)Defence articles & servicesAdministered by State Dept (DDTC); strict controls on military tech exports
Wassenaar Arrangement42 countries; conventional arms + dual-use techVoluntary; limits export of surveillance/intrusion tools
💡 Encryption Export: Strong encryption (AES-256) was historically restricted. Modern rules are relaxed for mass-market software but still require classification review.

Transborder Data Flow & Data Sovereignty

Transborder data flow: Movement of personal/sensitive data across national borders. Key pain points:
Data Sovereignty: Data is subject to the laws of the country where it is stored, not where the data subject resides
Data Localisation Laws: Some countries (Russia, China, India) require personal data to be stored on servers within the country
Transfer Mechanisms: GDPR requires adequacy decisions, SCCs, or BCRs for EU→third-country transfers (see D2 GDPR section)
Cloud Complication: Multi-region cloud storage means data may replicate to jurisdictions you didn’t intend — contractually restrict regions
Conflict-of-Laws: When two jurisdictions impose contradictory requirements (e.g., US CLOUD Act subpoena vs. GDPR data protection), seek legal counsel
⚠️ Exam Trap: Data sovereignty = governed by country where data is stored. Cloud provider’s headquarters location is irrelevant. Always know where your data physically resides.

1.6 Investigation Types

TypePurposeStandard of ProofWho Conducts
AdministrativeInternal policy violationPreponderanceInternal HR / Security
CriminalLaw enforcement prosecutionBeyond reasonable doubtLaw enforcement (FBI, police)
CivilDisputes / damagesPreponderance (51%+)Attorneys / courts
RegulatoryCompliance violationsSubstantial evidenceRegulatory bodies (SEC, GDPR DPA)
Industry StandardsPCI-DSS violationsContractual termsQSA (Qualified Security Assessor)
💡 Key: The burden of proof determines how strong evidence must be. Criminal investigations require the highest standard — therefore the strictest evidence handling.

1.7 Policies, Standards, Procedures & Guidelines

DocumentLevelNatureChangesExample
PolicyStrategic (WHY)Mandatory; signed by mgmtRare"All data must be encrypted at rest"
StandardTactical (WHAT)Mandatory; measurableModerate"Use AES-256 for encryption"
ProcedureOperational (HOW)Mandatory; step-by-stepFrequent"Click Settings → Enable encryption..."
BaselineMinimum configMandatory minimumPer platformCIS Benchmarks, DISA STIGs
GuidelineAdvisory (SHOULD)Optional; recommendedFlexible"Consider using password managers"

Policy Types

  • Regulatory: Required by law (HIPAA policy for healthcare orgs)
  • Advisory: Strongly recommended by management (AUP, clean desk)
  • Informative: Educational, no enforcement (security awareness newsletter)
🧠 Hierarchy Mnemonic — "Please Stop Playing Basketball, Gary": Policy → Standard → Procedure → Baseline → Guideline (most to least authoritative).
⚠️ Exam Trap: Policy = WHY (strategic, rarely changes). Standard = mandatory WHAT. Guideline = optional HOW. Procedure = mandatory step-by-step HOW. The exam LOVES testing this hierarchy.

1.8 Risk Management Concepts

Core Definitions

TermDefinition
AssetAnything of value
ThreatPotential harmful event
VulnerabilityExploitable weakness
RiskThreat × Vulnerability × Impact
ExploitMethod to leverage a vulnerability
ControlSafeguard/countermeasure to reduce risk
TermDefinition
Risk AppetiteRisk the org will accept (board sets)
Risk ToleranceAcceptable variation within appetite
Risk CapacityMax risk the org can absorb before failure
Residual RiskRisk remaining after controls
Inherent RiskRisk before any controls
Total RiskThreat × Vuln × Asset Value

Quantitative Risk Analysis — Core Formulas

SLE = AV × EF    (Single Loss Expectancy = Asset Value × Exposure Factor)
ALE = SLE × ARO    (Annualised Loss Expectancy = SLE × Annual Rate of Occurrence)
Value of Control = ALEbefore − ALEafter − Annual Cost of Control
ROSI = (ALEbefore − ALEafter − Cost) / Cost    (Return on Security Investment)
⚠️ EF is always a decimal: 50% damage = 0.50, not 50. ARO is frequency per year: 1 = once/year, 0.1 = once every 10 years, 3 = three times/year.

Qualitative Risk Analysis

  • Risk Matrix: Likelihood × Impact → colour-coded (Green/Yellow/Orange/Red)
  • Delphi Technique: Anonymous, iterative expert consensus — eliminates groupthink
  • OCTAVE: Asset-driven, qualitative (Carnegie Mellon)
💡 FAIR (Factor Analysis of Information Risk): Despite being listed alongside qualitative techniques for contrast, FAIR is a quantitative framework — uses Loss Event Frequency (LEF) × Loss Magnitude (LM).

Risk Response Strategies — "MATA"

ResponseActionResidual RiskExample
MitigateReduce via controlsReduced (still > 0)Install firewall, deploy MFA
AcceptDocumented decision to bear riskUnchangedLow-impact risk below appetite
TransferShift financial burden to third partyFinancial impact shiftedCyber insurance, outsourcing
AvoidEliminate the risky activityZero (only response = 0)Discontinue CC storage
🧠 "MATA" — Risk kills you if you don't respond. Mitigate · Accept · Transfer · Avoid. Only Avoid eliminates risk entirely.

Control Types Matrix

Function ↓ / Category →AdministrativeTechnicalPhysical
PreventivePolicies, screeningFirewalls, encryptionLocks, fences
DetectiveAudits, reviewsIDS, SIEM, logsCCTV, motion sensors
CorrectiveIncident response planPatching, AV quarantineFire suppression
DeterrentWarnings, AUPLogin bannersWarning signs, guards
RecoveryDRP, BCPBackups, failoverAlternate site
CompensatingSupervisionMonitoring when MFA unavailableEscort when badge fails
⚠️ Exam Trap: "Ignore" is NOT a valid risk response. The correct term is Accept (documented, conscious management decision). Transfer shifts financial burden but NOT legal accountability.

NIST RMF 7 Steps

🧠 "Please Can Sally Implement All Authorised Monitoring": Prepare → Categorise → Select → Implement → Assess → Authorise → Monitor

Continuous Improvement & Risk Maturity

Risk Maturity Levels: Ad Hoc (1) → Repeatable (2) → Defined (3) → Managed (4) → Optimised (5). Higher maturity = consistent, measured, continuously improving risk practices.
Continuous improvement requires: regular risk reassessment, updated threat landscape analysis, control effectiveness reviews, KRI/KPI tracking, and lessons learned integration.

1.9 Threat Modeling

STRIDE Framework (Microsoft)

ThreatViolatesControl
SpoofingAuthenticationMFA, certificates
TamperingIntegrityHashing, digital signatures
RepudiationNon-repudiationAudit logs, digital signatures
Information DisclosureConfidentialityEncryption, access controls
Denial of ServiceAvailabilityRate limiting, redundancy
Elevation of PrivilegeAuthorisationLeast privilege, sandboxing

Other Methodologies

MethodFocusKey Feature
PASTARisk-centric, 7 stagesBusiness impact alignment; attack simulation
DREADScoring (legacy)Damage · Reproducibility · Exploitability · Affected Users · Discoverability
Attack TreesAttacker goal decompositionRoot = goal, branches = paths; AND/OR nodes
MITRE ATT&CKReal-world TTPsTactic (why) → Technique (how) → Procedure (specific); for detection/hunting
LINDDUNPrivacy threatsLike STRIDE but for data privacy
DFD Elements: External Entities · Processes · Data Stores · Data Flows. Trust Boundaries are the critical security checkpoints between elements.
💡 PASTA = 7 Stages: Define Objectives → Technical Scope → App Decomposition → Threat Analysis → Vulnerability Analysis → Attack Modelling → Risk & Impact Analysis

1.10 Supply Chain Risk Management (SCRM)

Vendor Due Diligence (Strongest → Weakest)

#MechanismStrengthKey Point
1SOC 2 Type IIStrongest6–12 months operational effectiveness
2Right-to-Audit ClauseHighDirect assurance but expensive
3SOC 2 Type IMediumPoint-in-time snapshot only
4ISO 27001 CertificationMediumAnnual surveillance audits
5Penetration Test ResultsLow-MedTechnical only, no process assurance
6Security QuestionnaireWeakestSelf-reported, no verification

Supply Chain Threats

  • Build Pipeline Compromise: SolarWinds SUNBURST (2020) — malicious code in signed updates
  • Dependency Confusion: Attacker publishes higher-version malicious package in public repo
  • Typosquatting: Misspelled package names with malicious code
  • SBOM (Software Bill of Materials): Machine-readable inventory; mandated by US Executive Order 14028
  • Fourth-Party Risk: Your vendor's vendors — often uncontracted and invisible

Hardware Supply Chain Trust Mechanisms

MechanismDescription
Silicon Root of TrustHardware-anchored secure boot chain; cryptographic verification starts in immutable chip firmware — prevents firmware-level tampering
Physically Unclonable Function (PUF)Unique hardware fingerprint from manufacturing variations; used for device authentication and anti-counterfeiting

SLA Security Components

Uptime % · Incident Response Time · Breach Notification · Data Handling & Residency · Audit Rights · Subcontractor Requirements · Penalties

🔑 NIST SP 800-161: Primary SCRM guidance. Tiered approach. C-SCRM policy required for federal procurement.

1.11 Business Continuity Planning (BCP)

BCP vs. DRP

AspectBCPDRP
ScopeEntire business (people, processes, facilities, tech)IT systems only
GoalKeep operating during disruptionRestore systems after disruption
RelationshipParent planSubset of BCP

External Dependencies

BIA must identify external dependencies: Utility providers (power, water, telecom), cloud/SaaS providers, supply chain partners, ISPs, payment processors, DNS providers. A single external failure can cascade into full business disruption. Map dependencies and include them in RTO/RPO analysis.

BIA Metrics — "The Four Clocks"

MetricMeaningWho Sets ItKey Relationship
MTDMaximum Tolerable DowntimeBusiness (management)Deadline — business fails after this
RTORecovery Time ObjectiveIT + BusinessTime to restore system function
RPORecovery Point ObjectiveBusinessMax acceptable data loss (time)
WRTWork Recovery TimeITTime to verify & resume after restore
CRITICAL: RTO + WRT ≤ MTD   (If violated, plan is inadequate)

Availability & Reliability Metrics

MTBF = Mean Time Between Failures (higher = better reliability)
MTTR = Mean Time To Repair (lower = better recovery)
Availability = MTBF ÷ (MTBF + MTTR)

RPO Drives Backup Frequency

RPOBackup StrategyCost
24 hoursDaily backupLow
4 hoursEvery 4 hoursMedium
Near-zeroReal-time replicationHighest

Recovery Site Strategies

SiteActivationCostBest RTOIncludes
HotMinutes–hoursHighest< 4 hrsFull equipment + real-time data
WarmHours–daysMedium4–72 hrsPartial equipment + recent backups
ColdDays–weeksLowest> 72 hrsEmpty facility, power/cooling only
CloudMinutes–hoursPay-per-useFlexibleVirtual infrastructure on demand
MobileHoursMediumVariablePortable pre-configured container
ReciprocalVariableLowVariablePeer agreement — unreliable
🧠 HOT → WARM → COLD = READY → SOON → LATER. Temperature = urgency = expense.

BCP Testing Types (Least → Most Disruptive)

Test TypeDisruptionDescription
1. Checklist ReviewNoneReview plan document
2. Tabletop ExerciseNoneScenario walkthrough, discussion-based
3. SimulationLow-MedSimulated disaster, practice response
4. Parallel TestMediumActivate recovery site while primary runs
5. Full InterruptionHighShut down primary, switch to backup (most realistic)
🧠 "Can This Walk? Simulate Parallel Full!" Checklist → Tabletop → Walkthrough → Simulation → Parallel → Full Interruption

Vital Records & Plan Components

  • Vital Records: Legal, operational, emergency documents — store offsite in multiple copies
  • BCP document itself is a vital record — must be stored offsite
  • Plan Components "BCDCO-I": BCP (master) → DRP (IT) → Crisis Communications → OEP (evacuation) → COOP (govt) → IRP (cyber)
⚠️ FIRST priority in any disaster = PERSONNEL SAFETY. Always before system recovery.
⚠️ Exam Trap: Hot site ≠ always correct answer. Choose the cheapest option that meets MTD/RTO/RPO. Reciprocal agreements are unreliable — tempting wrong answer.

1.12 Personnel Security Policies

Employment Lifecycle — "HEAPS"

PhaseKey ActionsKey Controls
HireBackground checks, credential verification, credit historyPre-employment screening
Employment AgreementsNDA, NCA, AUP, IP Assignment signedLegal bindings established
Access ProvisioningLeast privilege + need to know + SoDRole-based access
PracticeJob rotation, mandatory vacationFraud detection + knowledge resilience
SeparationExit interview, equipment returnREVOKE ACCESS FIRST

Key Personnel Controls

ControlPurposeTypeKey Distinction
Separation of Duties (SoD)Prevent single-person fraudPreventiveTWO people needed simultaneously
Job RotationDetect fraud + build resilienceDetectivePeriodic REPLACEMENT of role
Mandatory VacationExpose fraud requiring presenceDetectiveTemporary REMOVAL (1-2 weeks)
Least PrivilegeMinimum permissions for dutiesPreventiveWHAT YOU DO (permissions)
Need to KnowInformation limited to necessityPreventiveWHAT YOU SEE (information)

Employment Agreements

AgreementPurposeDuration
NDAConfidentiality prohibitionSurvives termination
NCANon-compete restrictionTime & geo-limited (enforceability varies)
AUPIT resource usage rules; monitoring basisDuring employment; must sign before access
IP AssignmentWork product ownershipDuring employment scope
⚠️ FIRST action on termination = REVOKE ACCESS immediately — before HR meeting, exit interview, or equipment return. Especially for hostile terminations.
⚠️ SoD prevents fraud but NOT collusion. Multiple controls (audit, rotation, monitoring) needed for collusion detection.

1.13 Security Awareness, Education & Training

Three Tiers — "ATE"

TierAudienceDepthExamples
AwarenessALL staffShallow — recognitionPhishing posters, email tips, videos
TrainingRole-specific groupsSkills-based — measurableSecure coding for devs, BEC for finance
EducationSecurity professionalsDeep — conceptualCISSP, university programmes, SANS
🧠 "ATE": Awareness (Everyone) → Training (Role groups) → Education (Experts). Broader audience = shallower depth.

Social Engineering Vectors

VectorMethodDefence
PhishingMass deceptive emailAwareness + email filter + MFA
Spear PhishingTargeted individualRole training for high-value targets
WhalingC-suite targetingExecutive awareness + financial verification
VishingVoice call impersonationIdentity verification; callback protocol
SmishingSMS phishingMobile policy awareness
PretextingFalse scenario extractionNeed-to-know; verification protocols
TailgatingPhysical follow-throughBadge-only entry; anti-tailgating awareness
BaitingInfected USB dropPort controls; autorun blocking
🔑 Effectiveness Metrics: CISSP prefers behavioural/outcome metrics (phishing click rate, report rate) over activity metrics (completion rate, attendance). Measure what people do, not what they attended.

Modern Awareness Techniques

TechniqueDescription
Security ChampionsEmbedded advocates within dev/business teams; bridge between security team and staff; peer influence
GamificationPoints, leaderboards, badges for security behaviours; increases engagement over traditional CBT

Emerging Technology Awareness

Periodic content reviews must address: Cryptocurrency/blockchain risks (wallet theft, smart contract exploits), AI/LLM risks (prompt injection, deepfakes, data leakage), IoT threats, quantum computing impact on encryption, and social media privacy risks.
⚠️ Exam Trap: When asked about fixing human error, the answer is security training (addresses root cause), not just technical controls.

Domain 1 — All Formulas

Quantitative Risk:
SLE = AV × EF
ALE = SLE × ARO
Value of Control = ALEbefore − ALEafter − Annual Control Cost
ROSI = (ALEbefore − ALEafter − Control Cost) / Control Cost

BIA / Availability:
RTO + WRT ≤ MTD
Availability = MTBF ÷ (MTBF + MTTR)

Negligence (legal):
Duty → Breach → Causation → Damages (all 4 required)

Risk Equation:
Risk = Threat × Vulnerability × Impact

Domain 1 — Mnemonics Master List

TopicMnemonicExpands To
Ethics CanonsSAPASociety → Act honourably → Provide service → Advance profession
Policy Hierarchy"Please Stop Playing Basketball, Gary"Policy → Standard → Procedure → Baseline → Guideline
Risk ResponseMATAMitigate → Accept → Transfer → Avoid
BIA Metrics"The Four Clocks"MTD → RTO → RPO → WRT
NIST CSF 2.0GI-PD-RRGovern → Identify → Protect → Detect → Respond → Recover
NIST RMF"Please Can Sally Implement All Authorised Monitoring"Prepare → Categorise → Select → Implement → Assess → Authorise → Monitor
STRIDES-T-R-I-D-ESpoofing · Tampering · Repudiation · Information Disclosure · DoS · Elevation
BCP Testing"Can This Walk? Simulate Parallel Full!"Checklist → Tabletop → Walkthrough → Simulation → Parallel → Full Interruption
Recovery SitesHOT→WARM→COLDReady → Soon → Later (temp = urgency = cost)
Employment LifecycleHEAPSHire → Employ (agreements) → Access → Practice → Separation
Awareness TiersATEAwareness → Training → Education (wider = shallower)
Due Care vs DiligenceCare=Act, Diligence=CheckDue Care = doing right; Due Diligence = verifying
GDPR PrinciplesLPDA-SIALawfulness · Purpose · Data min · Accuracy · Storage · Integrity · Accountability
NegligenceD-B-C-DDuty → Breach → Causation → Damages
OECD Privacy8 principlesCollection · Quality · Purpose · Use · Security · Openness · Participation · Accountability

Domain 1 — Exam Traps & Common Mistakes

#TrapCorrect Answer
1Data Owner is IT / sysadminData Owner = business executive; Custodian = IT
2ISO 27002 is certifiableOnly ISO 27001 is certifiable; 27002 is guidance
3"Ignore" as risk responseNot valid. Use "Accept" (documented decision)
4Transfer eliminates accountabilityTransfers financial burden only, NOT legal liability
5Compliance = securityCompliance is the floor, not the ceiling
6Copyright protects ideasCopyright protects expression not ideas
7Trade secrets auto-protectedRequire active protection or they are lost
8GPL is safe for commercialGPL is viral/copyleft — derivative works must be GPL
9EF is a percentage number (50)EF is a decimal (0.50)
10Ransomware primarily hits confidentialityPrimary target = Availability (to extort payment)
11Hot site is always the right answerChoose cheapest option meeting MTD/RTO/RPO
12First action in disaster = restore systemsFirst action = personnel safety
13First action on termination = exit interviewFirst action = revoke access immediately
14SoD prevents collusionSoD prevents fraud, NOT collusion
15RTO = total recovery timeRTO + WRT ≤ MTD; WRT is verification time after restore
16Mandatory vacation is preventiveMandatory vacation = detective (exposes fraud)
17Training completion = effectivenessBehavioural metrics (click/report rate) > activity metrics
18Canon 3 (employer) over Canon 1 (society)Society always first — whistle-blowing justified
19Reciprocal agreements are reliableUnreliable — both orgs may face same disaster
20Least Privilege = Need to KnowLP = permissions (what you do); NtK = information (what you see)
21Due care = due diligenceDue care = doing the right thing (actions/implementation); due diligence = verifying it was done correctly
22CISO is ultimately responsible for securitySenior management is ultimately responsible; CISO advises
23Risk appetite = risk toleranceRisk appetite is strategic/broad; risk tolerance is tactical/specific
24Delphi technique = group brainstormingDelphi = anonymous expert consensus (reduces groupthink)
25BIA and Risk Assessment do the same thingBIA = impact; Risk Assessment = likelihood (both required)
26Criminal and administrative law are the sameCriminal/civil/administrative have different standards and penalties
27MTD is set by IT based on technical capabilityMTD is set by the business; IT aligns RTO/RPO to meet it
28Prudent person rule doesn't apply to cybersecurityPrudent Person Rule is used to assess due care in cybersecurity decisions

Domain 2: Asset Security

10% of Exam · Priority ★★☆☆☆

2.1 Asset Classification & Inventory

ITAM (IT Asset Management): Systematic process of tracking, managing, and optimising all IT assets throughout their lifecycle.

Classification Levels

GovernmentCommercial
Top SecretConfidential / Restricted
SecretPrivate
ConfidentialSensitive
Sensitive But Unclassified (SBU)Public
Unclassified
⚠ Exam Trap: "Confidential" is the LOWEST classified government level but the HIGHEST commercial level — a naming collision designed to confuse.
Classify UP: If a system processes data of multiple classifications, the entire system is classified at the highest level of data it handles.

2.2 Data Ownership Roles

RoleResponsibilityKey Point
Data OwnerSenior management; determines classification, access rules, protection requirementsULTIMATELY accountable; usually C-suite / VP
Data CustodianImplements & maintains technical controls (backups, encryption, access controls)Usually IT; follows Owner's directives
Data StewardEnsures data quality, metadata management, compliance with policiesDay-to-day data accuracy & integrity
System OwnerOwns the system processing the data; responsible for system securityDifferent from Data Owner
Data Controller (GDPR)Determines WHY and HOW personal data is processedPrimary GDPR liability
Data Processor (GDPR)Processes data on behalf of the controllerIf exceeds instructions → BECOMES controller
✓ Exam Tip: Owner = Accountability (business decision). Custodian = Implementation (technical execution). This distinction is HEAVILY tested.

2.3 Data Lifecycle (6 Stages)

StageKey Controls
1. CreateClassify at creation, assign owner, apply labels/tags
2. StoreEncryption at rest, access controls, backup strategy, geographic constraints
3. UseDRM/IRM enforcement, monitoring, DLP endpoint agents, TEE for sensitive ops
4. ShareEncryption in transit (TLS), DLP gateway, watermarking, access verification
5. ArchiveLong-term compliant storage, retention schedules, encryption key preservation
6. DestroyNIST 800-88 sanitisation (Clear/Purge/Destroy), Certificate of Destruction
🧠 Mnemonic: "Create Store Use Share Archive Destroy" = CSUSAD → "Cats Sleep Under Soft Autumn Duvets"

2.4 Data Handling & Marking Requirements

Marking = Making classification visible — headers/footers on documents, labels on media, metadata tags on digital assets, banner pages on printouts.
ClassificationHandling Requirements
Top Secret / ConfidentialEncrypted storage & transit, strict need-to-know, clean desk, secure destruction, audit logging
Secret / PrivateEncrypted where practical, role-based access, controlled distribution
Sensitive / InternalBasic access controls, labelling
PublicIntegrity protection (prevent unauthorised modification)

2.5 Privacy Protection Techniques

TechniqueDescriptionReversible?
AnonymisationIrreversibly removes all PII; cannot re-identify❌ No
PseudonymisationReplaces identifiers with tokens/aliases; can re-identify with mapping key✅ Yes
TokenisationSubstitutes sensitive value with non-sensitive token; token vault stores mapping✅ Yes (with vault)
Data MaskingHides portions of data (e.g., XXX-XX-1234); static or dynamicDepends on type
GeneralisationReduces precision (exact age → age range)❌ No
Privacy by Design — 7 Principles (Cavoukian):
1. Proactive not Reactive · 2. Privacy as Default · 3. Privacy Embedded in Design · 4. Full Functionality (positive-sum) · 5. End-to-End Security · 6. Visibility & Transparency · 7. Respect for User Privacy
⚠ Exam Trap: Anonymisation = GDPR no longer applies (not personal data). Pseudonymisation = GDPR still applies (can be re-identified).

2.6 Data Retention & Destruction

Retention Periods

RegulationRetention
SOX (Financial)7 years
HIPAA (Health)6 years
PCI-DSS (Payment)1 year
GDPRAs long as necessary (minimisation)
⚠ Litigation Hold: Overrides ALL retention schedules — if litigation is reasonably anticipated, STOP all destruction immediately.

NIST SP 800-88 Sanitisation Levels

LevelMethodRobustnessMedia Reusable?
ClearLogical overwrite (patterns)Defeats standard recoveryYes
PurgeDegaussing, ATA Secure Erase, crypto-eraseDefeats forensic labUsually
DestroyShredding, incineration, disintegrationMedia non-functionalNo
🧠 "CPD — Can Parrots Dance?" — Clear → Purge → Destroy (escalating permanence)

Critical Degaussing Rules

Degaussing works on: ✅ Spinning HDDs, ✅ Magnetic tape
Degaussing has NO effect on: ❌ SSDs, ❌ Flash drives, ❌ NVMe, ❌ Optical (CD/DVD), ❌ Smartphones, ❌ SD cards
⚠ Exam Trap: Degaussing SSDs = ZERO effect. DoD 5220.22-M overwrite is UNRELIABLE for SSDs. Preferred SSD purge = crypto-erase or ATA Secure Erase Enhanced.

Cloud Data Destruction

✓ Only practical option: Cryptographic erasure — encrypt data before upload with YOUR key; destroy the key → all copies (replicas + backups) become crypto-garbage.

Data Remanence

Data Remanence: Residual data left on storage media after deletion or formatting. Standard file deletion only removes pointers, not actual data. Even after overwriting, magnetic/electric traces may remain on some media. Remanence is why sanitisation (NIST 800-88) exists — simple deletion is NEVER sufficient for sensitive data.

End-of-Life (EOL) & End-of-Support (EOS)

TermMeaningAction Required
EOLVendor stops selling/manufacturingPlan migration; no new deployments
EOSVendor stops patches & supportReplace or isolate; no security updates → critical risk
⚠ EOS systems = unpatched = high-risk. If replacement impossible: isolate (VLAN/firewall), increase monitoring, add compensating controls, document risk acceptance.

2.7 Data Security Controls

Scoping & Tailoring

Scoping: Selecting which controls from a baseline apply to a specific system based on its environment and technology.
Tailoring: Customising selected controls to fit the organisation’s specific risk, mission, and operational requirements.
Scope first (include/exclude), then tailor (adjust parameters).

Data States & Protection

StateWhereProtection
At RestDisk, DB, tape, SANAES-256, FDE, TDE, file-level encryption
In TransitNetwork, Internet, VPNTLS 1.3, IPSec, SSH, HTTPS
In UseCPU, RAM, registersTEE (SGX, TrustZone), homomorphic encryption
🧠 "RAT": Rest (AES/FDE), Air/transit (TLS/IPSec), Thinking/use (TEE/SGX)

Key Security Technologies

TechnologyPurpose
DRM/IRMPersistent content protection; controls copy, print, forward, screenshot
CASBCloud security broker: Visibility, Compliance, Data Security, Threat Protection
DLPContent-based detection & blocking of unauthorised data flows
TPMHardware crypto processor on motherboard; stores keys, integrity measurements
HSMDedicated hardware for key management; FIPS 140-2/3 Level 3+; tamper-resistant
⚠ TPM vs HSM: TPM = soldered on motherboard, single-device. HSM = external appliance/network device, enterprise-scale. Both store keys in hardware.

2.8 Data Loss Prevention (DLP)

DLP TypeLocationData StateKey Threat
Network DLPGateway / proxyIn transitEmail/web exfiltration
Endpoint DLPAgent on deviceAt rest + in useUSB copy, print, off-network
Cloud DLP / CASBAPI to cloudAt rest in cloudShadow IT, SaaS exposure
DLP Deployment Sequence: Monitor → Alert → Block (never jump to block — causes alert fatigue).
✓ Content Inspection: Regex (fast, high FP), Fingerprinting (exact match, low FP), Statistical/ML (unstructured), Contextual (who + when + where).
⚠ Network DLP requires SSL/TLS inspection to see inside HTTPS traffic. Without it, all encrypted traffic is opaque.

2.9 GDPR Framework

7 GDPR Principles (Article 5)

L-P-D-A-S-I-A: Lawfulness · Purpose limitation · Data minimisation · Accuracy · Storage limitation · Integrity & Confidentiality · Accountability

Data Subject Rights (8 Rights)

🧠 "I ACCESS DP": Informed · Access · Correction · Curtail (restrict) · Erasure · Switching (portability) · Stop (object) · Decision-making · Profiling

Key GDPR Rules

TopicRule
Max Fine€20M or 4% global annual turnover (whichever higher)
Breach → Supervisory AuthWithin 72 hours of controller awareness
Breach → Data SubjectsWithout undue delay IF high risk
DPO Required WhenPublic body, large-scale monitoring, special category data
Right to ErasureNOT absolute — 6 exceptions (legal claims, legal obligation, public health, archiving, freedom of expression, public interest)

Cross-Border Transfer Mechanisms

MechanismDescription
Adequacy DecisionEC determines country has adequate protection
SCCsPre-approved contractual clauses (most widely used post-Schrems II)
BCRsIntra-group rules for multinationals (complex, expensive)
EU-US DPFJuly 2023 adequacy decision (replacing Privacy Shield)
⚠ Schrems II (2020): Invalidated Privacy Shield. SCCs now require Transfer Impact Assessments (TIAs).
⚠ DPO Conflict Trap: DPO cannot simultaneously be CISO, IT Director, CEO, or any role determining processing purposes.

Domain 2 — Mnemonics

#MnemonicHelps Remember
1"Cats Sleep Under Soft Autumn Duvets"Data Lifecycle: Create, Store, Use, Share, Archive, Destroy
2"CPD — Can Parrots Dance?"Sanitisation: Clear → Purge → Destroy
3"RAT"Data states: Rest, Air/transit, Thinking/use
4"NEC"DLP types: Network, Endpoint, Cloud/CASB
5"MAB"DLP sequence: Monitor, Alert, Block
6"SSD = Sad, Stubborn Data"Overwriting doesn't work, degaussing worthless; use crypto-erase
7"Cloud Key = Cloud Bye"Destroy encryption key = destroy all cloud copies
8"I ACCESS DP"GDPR data subject rights
9"L-P-D-A-S-I-A"GDPR 7 principles
10"72 → SA; High Risk → Subject"GDPR breach notification timeline

Domain 2 — Exam Traps 🚨

#Trap
1"Confidential" is LOWEST govt classified but HIGHEST commercial level
2Data Owner = senior management (NOT IT staff); Custodian = IT
3Degaussing SSDs = ZERO effect (data fully intact)
4DoD 5220.22-M overwrite UNRELIABLE for SSDs (wear-leveling bypasses it)
5Crypto-erase = preferred for SSDs AND cloud
6Degaussed HDD becomes permanently inoperable (servo tracks destroyed)
7Litigation hold OVERRIDES all retention schedules
8Anonymised data = GDPR no longer applies; Pseudonymised = GDPR still applies
9Processor exceeding instructions BECOMES a controller (full liability)
10Right to Erasure is NOT absolute (6 major exceptions)
11DPO cannot hold CISO/CTO/CEO roles simultaneously (conflict of interest)
1272-hour GDPR clock starts at controller awareness, not incident discovery
13Network DLP requires SSL inspection to see HTTPS content
14Data in use is HARDEST to protect (requires hardware TEE)
15FDE useless on powered-on, logged-in devices
16FALSE: Data steward = data owner. CORRECT: Data steward manages data quality, metadata, and day-to-day governance. Data owner (senior management) determines classification and access policy — they are ultimately accountable for the data.
17FALSE: Scoping = tailoring. CORRECT: Scoping = selecting which baseline controls apply to a specific system (include/exclude). Tailoring = customising the parameters of chosen controls to fit the organisation. Scope first, then tailor.
18FALSE: Marking and labelling are the same. CORRECT: Marking = making classification visible on document content (headers, footers, banner pages, metadata tags). Labelling = physical indicator on media or hardware (adhesive label, barcode tag). Marking is for documents; labelling is for physical assets.
19FALSE: All sanitisation methods are equivalent. CORRECT: NIST SP 800-88 defines three escalating levels — Clear (logical overwrite; media reusable; defeats standard recovery tools) → Purge (degauss / crypto-erase / secure-erase; defeats forensic tools; media often reusable) → Destroy (shred/incinerate; no reuse; highest assurance). Match level to data sensitivity and media reuse requirement.

Domain 3: Security Architecture & Engineering

13% of Exam · Priority ★★★☆☆

3.1 Security Design Principles (Saltzer & Schroeder)

PrincipleMeaning
Least PrivilegeMinimum access needed to perform function
Economy of MechanismKeep security simple; complex = more bugs
Separation of PrivilegeRequire multiple conditions/people (dual control)
Separation of DutiesNo single person controls entire critical process
Complete MediationCheck EVERY access, not just first (no caching authz)
Open DesignSecurity not dependent on secrecy of design (Kerckhoffs')
Psychological AcceptabilitySecurity must not make system unusable
Fail-Safe DefaultsDefault = deny access; explicitly grant
🧠 "LESS COPS": Least privilege · Economy · Separation (privilege) · Separation (duties) · Complete mediation · Open design · Psychological acceptability · Safe defaults

3.2 Formal Security Models

Bell-LaPadula (BLP) — Confidentiality

Purpose: Protects CONFIDENTIALITY in multi-level security (MLS) systems.
Simple Security (ss): No Read Up — subject cannot read object at higher clearance
Star Property (*-property): No Write Down — subject cannot write to object at lower clearance
Strong Star: Read + Write only at own level
🧠 "WURD": Write Up, Read Down (what BLP allows) — OR "No Read Up, No Write Down"

Biba — Integrity

Purpose: Protects INTEGRITY (opposite of BLP).
Simple Integrity: No Read Down — don't read from lower integrity (no contamination)
Star Integrity (*): No Write Up — don't write to higher integrity (no corruption)

Clark-Wilson — Commercial Integrity

CDI: Constrained Data Items (protected data)
TP: Transformation Procedures (authorised operations on CDIs)
IVP: Integrity Verification Procedures (validate CDI consistency)
Enforces well-formed transactions and separation of duties.

Other Models

ModelPurposeKey Feature
Brewer-Nash (Chinese Wall)Conflict of interestAccess rules change dynamically based on access history
Graham-DenningObject management8 basic operations (create/delete subject/object, read/grant/delete/transfer access)
HRUAccess control decidabilitySafety problem is UNDECIDABLE in general case
Take-GrantAccess rights transfer4 operations: take, grant, create, remove
LipnerCombined modelUses BLP + Biba together for both C & I
⚠ Exam Trap: BLP ↔ Biba are EXACT opposites. BLP = confidentiality = no read up/write down. Biba = integrity = no read down/write up. Confusing their rules is the #1 trap.
💡 See also: Section 5.4 Access Control Models (D5) — DAC, MAC, RBAC, and ABAC are the practical enforcement mechanisms arising from these theoretical security models. The BLP and Biba models map directly to MAC enforcement in real systems.

3.3 Security Architecture Frameworks

FrameworkFocusKey Feature
ZachmanEnterprise architecture6×6 taxonomy matrix (What/How/Where/Who/When/Why × perspectives)
TOGAFEnterprise architectureArchitecture Development Method (ADM) — iterative cycle
SABSASecurity-specific architecture6 layers aligned to business risk; security-focused (unlike Zachman/TOGAF)
✓ Exam Tip: SABSA = Security Architecture. Zachman & TOGAF = Enterprise Architecture (not security-specific).

3.4 Security Evaluation Criteria

StandardOriginLevelsStatus
TCSEC (Orange Book)US DoDD → C1 → C2 → B1 → B2 → B3 → A1Historical / Legacy
ITSECEuropeE0 → E6 (+ F ratings for functionality)Historical / Legacy
Common Criteria (CC)International (ISO 15408)EAL 1–7✅ CURRENT standard

Common Criteria Components

PP (Protection Profile): Customer requirements ("what the community needs")
ST (Security Target): Vendor's claims ("what the product does")
TOE (Target of Evaluation): The actual product being evaluated
🧠 "PP = People Profile; ST = Seller Target; TOE = Thing Evaluated"

EAL Levels Quick Reference

EALNameKey Point
1Functionally TestedLowest assurance
2Structurally TestedBasic analysis
3Methodically TestedEngineering discipline
4Methodically Designed, Tested, ReviewedHighest commercially practical
5Semi-formally Designed & TestedSignificant expense
6Semi-formally VerifiedHighly specialised
7Formally VerifiedFormal proof; highest assurance

3.5 System Security Architecture

TCB (Trusted Computing Base): All hardware, firmware, and software responsible for enforcing security policy. KEEP IT SMALL (smaller = fewer bugs = more verifiable).
Reference Monitor: Abstract concept — mediates ALL access between subjects and objects. Properties: Complete mediation, Tamper-proof, Verifiable (small enough to verify).
Security Kernel: The actual implementation of the reference monitor concept in hardware/software/firmware.
✓ Tip: Reference Monitor = concept. Security Kernel = implementation. TCB = everything enforcing policy.

3.6 CPU & Memory Security

Protection Rings

RingLevelWhat Runs Here
Ring -1HypervisorType 1 hypervisor (VMM) below kernel
Ring 0KernelOS kernel — full hardware access
Ring 1–2Drivers / ServicesDevice drivers, OS services
Ring 3UserApplications — most restricted

Memory Protections

ControlPurpose
ASLRRandomises memory layout; defeats predictable exploits
DEP/NXMarks memory pages non-executable; prevents code injection from running
Stack CanarySentinel value detects buffer overflow before return

3.7 Virtualisation & Cloud Security

Virtualisation

TypeDescriptionExample
Type 1 (Bare-Metal)Runs directly on hardwareVMware ESXi, Hyper-V, Xen
Type 2 (Hosted)Runs on top of host OSVirtualBox, VMware Workstation
⚠ VM Escape: Attack where malicious code breaks out of VM into hypervisor or host. Most critical virtualisation threat.

VM vs Container

AspectVMContainer
IsolationFull (own OS kernel)Shared kernel (weaker)
OverheadHigh (full OS per VM)Light (shared OS)
Boot timeMinutesSeconds
SecurityStronger isolationKernel compromise = all containers

Cloud Service Models (Shared Responsibility)

ModelCustomer ControlsProvider Controls
IaaSOS, apps, data, middleware, runtimeHardware, network, virtualisation
PaaSApps, data+ OS, middleware, runtime
SaaSData, some configEverything else
CASB 4 Pillars: Visibility · Compliance · Data Security · Threat Protection

3.8 IoT & Embedded / ICS Systems

⚠ SCADA/ICS Priority: Safety > Availability > Integrity > Confidentiality (SAIC, not CIA!)
ComponentRole
SCADASupervisory control; wide-area monitoring
PLCProgrammable Logic Controller; controls physical processes
RTURemote Terminal Unit; field data collection
HMIHuman-Machine Interface; operator dashboard
DCSDistributed Control System; local plant automation
✓ IoT Threats: Default credentials, no encryption, no patching, Mirai botnet (2016 DDoS via default-password IoT devices).

Additional Architecture Types

ArchitectureSecurity Concern
Serverless (FaaS)No OS to manage; provider controls runtime; risks = function injection, excessive permissions, insecure dependencies, cold start timing attacks
Edge ComputingProcessing at network edge (near data source); limited physical security; constrained resources for encryption; supply chain risk for edge devices
High-Performance Computing (HPC)Massive parallel processing; risk = shared memory/interconnects; data classification challenges; multi-tenant GPU clusters
Microservices / APIDecomposed applications; each service = attack surface; requires API gateway, mutual TLS, and rate limiting between services

3.8b Site & Facility Design

CPTED (Crime Prevention Through Environmental Design): Design physical spaces to deter crime through 3 strategies:
StrategyDescriptionExample
Natural SurveillanceDesign that maximises visibilityLow hedges, open sight lines, windows facing parking
Natural Access ControlGuide movement through designSingle entry point, pathways, landscaping barriers
Territorial ReinforcementMark ownership boundariesSignage, fences, lighting, maintained landscaping

Fence Heights (Classic Exam Topic)

HeightPurpose
3–4 ft (1 m)Deters casual trespass; defines boundary
6–7 ft (2 m)Too high to climb easily; deters most intruders
8 ft + barbed/razor wireDeters determined intruders; critical areas

Facility Secure Areas

AreaSecurity Requirements
Wiring Closets / IDFLocked; restricted access; no storage of personal items; environmental monitoring
Server Rooms / Data CentresMulti-factor entry; CCTV; HVAC; fire suppression; visitor logs; raised floors
Media Storage FacilitiesFireproof, climate-controlled; access restricted to custodians; inventory tracking
Evidence StorageTamper-evident; strict chain of custody; limited access; environmental controls

Power Issues

EventDescriptionProtection
BlackoutTotal power lossUPS (short-term) + Generator (long-term)
BrownoutProlonged low voltageUPS with voltage regulation
Surge/SpikeExcess voltage (spike = instant, surge = sustained)Surge protector / UPS
SagMomentary low voltageUPS / line conditioner
NoiseEMI/RFI interferenceLine conditioners, shielded cabling
InrushInitial surge when power restoredUPS / managed power-on sequence
✓ UPS Types: Online (double conversion) = best protection, always on battery; Standby = cheapest, switches on outage; Line-interactive = middle ground.

HVAC & Environmental

FactorIdeal RangeRisk
Temperature64–75°F / 18–24°COverheating → component failure
Humidity40–60% RHToo low → ESD; too high → condensation/corrosion
⚠ Water Detection: Install water/moisture sensors under raised floors & near HVAC. Positive pressurisation prevents dust/contaminants.

Facility Location Selection

  • Avoid flood plains, earthquake zones, flight paths, chemical plants
  • Consider proximity to emergency services (fire, police, hospital)
  • Low visibility preferred — avoid rooftop signs, public identification
  • Redundant ISP entry points from different directions

3.9 Cryptography Fundamentals

Kerckhoffs' Principle: Security depends on the KEY, not secrecy of the algorithm.
ConceptDefinition
ConfusionComplex relationship between key → ciphertext (substitution)
DiffusionSpreading plaintext across ciphertext (transposition/permutation)
Work FactorEffort/time to break a cipher (key space = 2^n)

3.10 Symmetric Encryption

AlgorithmBlockKey SizeRoundsStructureStatus
DES645616Feistel❌ Broken
3DES64112/16848Feistel×3⚠️ Deprecated
AES128128/192/25610/12/14SPN✅ Standard
Blowfish6432-44816FeistelLegacy
Twofish128128-25616Feistel varSecure (not standardised)

Block Cipher Modes

ModeIV?Secure?Auth (AEAD)?Parallelisable?
ECBNo❌ INSECUREBoth
CBCYesDecrypt only
CTRNonceBoth (stream-like)
GCMNonceBoth
⚠ ECB penguin: Identical plaintext blocks → identical ciphertext blocks (patterns leak). NEVER use ECB.
🧠 "GCM = Gets both Confidentiality & Message auth" / "DES = 56-64-16" / "AES block = 128 ALWAYS"

3.11 Asymmetric Encryption

AlgorithmHard ProblemEncryptSignKey ExchMin Key
RSAInteger factorisation2048
DHDiscrete log2048
DSADiscrete log2048
ElGamalDiscrete log2048
ECCEC discrete log256

Key Count Formulas

Symmetric keys needed (n users): n(n−1) / 2
Asymmetric keys needed (n users): 2n
Key Size Equivalence: ECC-256 ≈ RSA-3072 ≈ 128-bit security level
Perfect Forward Secrecy (PFS): Use ephemeral DH (DHE/ECDHE). Past sessions safe even if long-term key compromised. TLS 1.3 MANDATES PFS.

Hybrid Cryptosystem (TLS)

Key exchange = asymmetric (slow, ECDHE) → Bulk data = symmetric (fast, AES-GCM)

3.12 Hashing & Digital Signatures

Hash Algorithms

AlgorithmOutputStatus
MD5128-bit❌ Broken (2004)
SHA-1160-bit❌ Broken (SHAttered, 2017)
SHA-256256-bit✅ Standard (Merkle-Damgård)
SHA-3 (Keccak)224-512✅ Standard (Sponge construction)
Birthday Attack Complexity: 2n/2 (for n-bit hash)
128-bit hash → only 264 operations to find collision

What Provides Non-Repudiation?

MechanismNon-Repudiation?Why
EncryptionNo proof of origin
HashingNo key = no identity
HMACShared key — both could generate
Digital SignaturePrivate key is unique to signer

Password Hashing Best Practices

Salt + Pepper + Stretch:
Salt: Unique random per password (defeats rainbow tables)
Pepper: System-wide secret stored separately from DB
Stretch: Slow functions — bcrypt, scrypt, Argon2 (make brute-force expensive)

3.13 PKI & Certificate Management

ComponentRole
CAIssues & signs certificates
RAVerifies identity (does NOT sign)
CRLPeriodic list of revoked cert serial numbers
OCSPReal-time cert status: good/revoked/unknown

CRL vs OCSP

FeatureCRLOCSP
MechanismDownload listReal-time query
FreshnessHours/days (stale)Current
PrivacyNo concernResponder sees queries
FixDelta CRLsOCSP Stapling

Trust Models

ModelAuthorityUse Case
HierarchicalRoot CA → Sub-CAsTLS/HTTPS, enterprise PKI
Web of TrustDecentralised peer-signedPGP/GPG email
Bridge CANeutral connectorFederal PKI
MeshCAs cross-certifyM&A / partnerships
Key Escrow Rule: ONLY escrow encryption keys (for data recovery). NEVER escrow signing keys (destroys non-repudiation). Use M-of-N splitting.

Post-Quantum Cryptography

Quantum Key Distribution (QKD): Uses quantum mechanics (photon polarisation) to distribute keys; eavesdropping disturbs quantum states and is detectable. Ensures theoretically unbreakable key exchange.
Post-Quantum Algorithms: NIST selected lattice-based (CRYSTALS-Kyber for KEM, CRYSTALS-Dilithium for signatures) as PQ standards. "Harvest now, decrypt later" = adversaries store encrypted data today to decrypt when quantum computers mature.

3.14 Cryptographic Attacks & Key Management

Attack Hierarchy (weakest → strongest)

AttackAttacker HasExample
Ciphertext-Only (COA)Only ciphertextFrequency analysis
Known-Plaintext (KPA)Some plaintext-ciphertext pairsKnown file headers
Chosen-Plaintext (CPA)Can encrypt arbitrary plaintextsBEAST attack
Chosen-Ciphertext (CCA)Can decrypt arbitrary ciphertextsMost powerful

Side-Channel Attacks

AttackExploitsCountermeasure
TimingExecution time variesConstant-time operations
Power AnalysisPower consumptionPower filtering, HSM
EM AnalysisElectromagnetic emissionsShielding, Faraday cage
Fault InjectionInduced errorsFault detection
✓ Side-channel attacks exploit the IMPLEMENTATION, not the algorithm math.

Additional Cryptanalytic Attack Types

AttackDescriptionCountermeasure
Pass the HashAttacker captures password hash and replays it to authenticate without crackingCredential Guard, privileged access workstations, Kerberos AES
Kerberos ExploitationGolden Ticket (forged TGT), Silver Ticket (forged service ticket), Kerberoasting (offline cracking)Protect KRBTGT, use AES, short ticket lifetimes, PAM
Frequency AnalysisAnalyse ciphertext letter frequencies to break substitution ciphersUse modern ciphers with diffusion (not simple substitution)
Brute ForceTry all possible keys exhaustivelyLonger keys, account lockout, rate limiting
MITM (Crypto)Intercept key exchange; present own keys to both partiesCertificate pinning, mutual authentication, PKI

Key Management 7 Phases

Generate → Distribute → Store → Use → Rotate → Archive → Destroy
Archive encryption keys (old data). NEVER archive signing keys. Use CSPRNG + FIPS 140-2/3 validated devices.

Key Stretching Algorithms

AlgorithmBest For
PBKDF2NIST recommended; CPU-intensive iterations
bcryptBlowfish-based; configurable cost factor
scryptCPU + memory intensive; resists GPU/ASIC
Argon2Current best practice; configurable time/memory/parallelism
Split Knowledge: Key divided; no one person knows full key.
Dual Control: Two+ people must collaborate to perform operation.

Information System Lifecycle (CBK 3.10)

PhaseSecurity Activity
Stakeholder NeedsDefine security requirements alongside business requirements
Requirements AnalysisSecurity classification, threat modelling, compliance requirements
Architectural DesignSecure architecture patterns, defence in depth, trust boundaries
Development / ImplementationSecure coding, code review, SAST
IntegrationInterface testing, system integration testing, DAST
Verification & ValidationSecurity testing, certification evaluation, acceptance testing
Transition / DeploymentHardening, configuration baseline, change management
Operations & MaintenanceContinuous monitoring, patching, incident response
Retirement / DisposalData sanitisation (NIST 800-88), media destruction, licence decommission

3.15 Security Modes of Operation (System Clearance Modes)

ModeClearance RequiredNeed-to-KnowAccessDescription
DedicatedALL users cleared to HIGHEST levelALL users have NtK for ALL dataAll dataSingle-classification processing; everyone cleared for everything
System HighALL users cleared to HIGHEST levelNOT all users have NtK for all dataBased on NtKAll cleared, but access restricted by need-to-know
CompartmentedALL users cleared to HIGHEST levelNtK + formal approval for compartmentsCompartment-basedAdds compartment approval beyond clearance + NtK
MultilevelNOT all users cleared to highestNtK requiredLabel-based (MAC)Users at different clearance levels; BLP/Biba enforced by system; MOST complex
🧠 "Does She Cook Meals?" — Dedicated → System High → Compartmented → Multilevel (increasing complexity, decreasing trust assumptions)
⚠️ Exam Trap: Dedicated = simplest (everyone cleared for everything). Multilevel = most complex (different clearances coexist). System High vs Compartmented: both require full clearance, but compartmented adds formal compartment approval.

3.16 Covert Channels

Covert Channel: Unintended communication path that violates security policy — transfers information in ways not designed or monitored.
TypeMechanismExampleDetection
Covert Storage ChannelOne process writes, another reads shared storageModifying file attributes, disk space, shared memory flagsAudit logs, resource monitoring
Covert Timing ChannelSignal via timing of system operationsCPU utilisation patterns, packet timing, lock contentionHarder to detect; noise injection, traffic normalisation
⚠️ Timing channels are HARDER to detect and eliminate than storage channels. Both are BLP violations (information flows down without write-down).
💡 Covert channels bypass mandatory access controls. They are a key concern in systems using BLP/multilevel security. TCSEC B2 requires covert storage channel analysis; B3 adds covert timing channel analysis and mitigation.

3.17 Additional Security Model Concepts

State Machine Model

State Machine Model: System is always in a defined "state" (snapshot of all permissions, data, and configurations). Secure if: every possible state transition preserves the security policy. If any reachable state violates policy → INSECURE.

Information Flow Model

Information Flow Model: Controls how information moves between subjects and objects at different security levels. BLP and Biba are both information flow models. Prevents both overt and covert flows that violate policy.

Lattice-Based Access Control

Lattice Model: Mathematical structure defining upper and lower bounds of access. Each subject and object has a security label. Access is allowed only if the subject's label dominates the object's label. Provides the mathematical foundation for MAC. BLP uses a lattice of classifications + compartments.

Process Isolation

MechanismDescription
Process IsolationOS prevents one process from accessing another's memory space; fundamental to multi-user security
Hardware SegmentationMemory segments enforced by MMU; stronger than software isolation alone
Virtual Address SpaceEach process sees its own virtual memory; OS maps to physical addresses; prevents cross-process access
SandboxingRestricts process to limited resource set; used for untrusted code (browser JS, mobile apps)
TEE (Trusted Execution Environment)Hardware-isolated enclave for sensitive operations (Intel SGX, ARM TrustZone); code + data protected even from OS

3.18 Memory Types & Architecture

TypeCategoryVolatile?Key Characteristic
SRAMRAMYesFast (L1/L2 cache); expensive; no refresh needed
DRAMRAMYesMain memory; cheaper; needs constant refresh
ROMFirmwareNoRead-Only; burned at factory; cannot be changed
PROMFirmwareNoProgrammable ROM; write-once (fuse-based)
EPROMFirmwareNoErasable PROM; UV light erases; rewritable
EEPROMFirmwareNoElectrically erasable; byte-level rewrite; slower
FlashFirmware/StorageNoType of EEPROM; block-level erase; SSDs, USB drives, firmware
⚠️ Security Concern: Volatile memory (RAM) loses data on power-off — collect FIRST in forensics. Non-volatile memory (ROM/flash) retains data — firmware rootkits can persist here.
💡 Cold Boot Attack: RAM doesn't lose data instantly on power-off. Attacker can freeze RAM chips (canned air) and extract encryption keys. Countermeasure: memory encryption, rapid key zeroisation.

3.19 Cloud Deployment Models

ModelOperated ByFor WhomKey Concern
PublicCloud providerGeneral public / any tenantMulti-tenancy; shared infrastructure; data residency
PrivateOrganisation or third partySingle organisation onlyHigher control; higher cost; still need security config
HybridMix of public + privateOrganisationData classification drives placement; policy consistency across environments
CommunityShared by orgs with common interestsMembers of community (govt, healthcare)Shared compliance requirements; FedRAMP community cloud
Multi-CloudMultiple public providersOrganisationAvoid vendor lock-in; complexity of managing security across providers
⚠️ Exam Trap: Private cloud ≠ on-premises. A private cloud can be hosted off-site by a third party — key is exclusive use by one organisation. Community cloud = shared by orgs with similar regulatory needs (e.g., GovCloud).

Domain 3 — Key Formulas & Numeric Facts

Key Counts:
Symmetric keys needed (n users): n(n−1) / 2
Asymmetric keys needed (n users): 2n (n public + n private = n key pairs)
Example: 10 users → Symmetric = 45 keys; Asymmetric = 20 keys (10 public + 10 private = 10 key pairs)

Hash Collision Resistance (Birthday Attack):
Complexity to find collision in n-bit hash: 2n/2
SHA-256 (256-bit) → 2128 operations to find collision (safe)
SHA-1 (160-bit) → 280 → broken 2017 (SHAttered); MD5 (128-bit) → 264 → broken 2004

Encryption Work Factor:
Brute-force key search space: 2n (for n-bit key)
DES 56-bit: 256 ≈ 72 quadrillion operations — cracked in <1 day with modern hardware
AES-128: 2128 — computationally infeasible; AES-256: 2256 — quantum-safe

Block Cipher Fixed Facts:
AES block size: always 128 bits (key = 128 / 192 / 256 bits)
DES: 56-bit key · 64-bit block · 16 rounds · Feistel structure
3DES effective key: 112-bit (2-key EDE) or 168-bit (3-key EDE) — deprecated

Key Size Equivalence (Approximate Security Level):
ECC-256 ≈ RSA-3072 ≈ DSA-3072 ≈ 128-bit symmetric ≈ 128-bit security level
ECC-384 ≈ RSA-7680 ≈ 192-bit security level
ECC-521 ≈ RSA-15360 ≈ 256-bit security level

EAL Levels (Common Criteria):
EAL 1–4: commercially achievable range · EAL 4 = highest commercially practical
EAL 5–7: require semi-formal or formal proofs — rarely achieved outside government/military

Domain 3 — Mnemonics

#MnemonicHelps Remember
1"LESS COPS"Saltzer & Schroeder 8 principles
2"WURD" (Write Up Read Down)What BLP allows (opposite = what it blocks)
3"PP = People Profile; ST = Seller Target; TOE = Thing Evaluated"Common Criteria components
4"GCM = Gets Confidentiality & Message auth"Only AEAD mode on the exam
5"DES = 56-64-16"DES key-block-rounds
6"AES block = 128 ALWAYS"Key varies (128/192/256) but block never changes
7"RSA Does It All (E/D/K)"Encrypt, Digital sign, Key exchange
8"ECC Punches Above Its Weight"ECC-256 ≈ RSA-3072
9"Sym = n(n−1)/2; Asym = 2n"Key count formulas
10"MD5 & SHA-1 are DEAD; SHA-2 = standard; SHA-3 = spare"Hash algorithm status
11"Salt Pepper Stretch"Password hashing best practices
12"Never escrow signatures"Key escrow rule
13"CPTED = See Naturally"Natural Surveillance, Natural Access, Territorial Reinforcement
14"Blackout = Both (UPS + Gen)"Total loss needs short-term + long-term power
15"Does She Cook Meals?"Security modes: Dedicated → System High → Compartmented → Multilevel
16"Storage = Stuff; Timing = Tempo"Covert channel types
17"ROM → PROM → EPROM → EEPROM → Flash"Memory permanence evolution (fixed → reprogrammable)

Domain 3 — Exam Traps 🚨

#Trap
1BLP = confidentiality (no read up/write down); Biba = integrity (no read down/write up) — OPPOSITES
2Clark-Wilson enforces well-formed transactions via TPs on CDIs (commercial integrity)
3Brewer-Nash = dynamic rules based on access HISTORY (not static)
4SABSA = security framework; Zachman/TOGAF = enterprise architecture (not security-specific)
5EAL 4 = highest practically achievable; EAL 7 = formal verification (extremely rare)
6Reference Monitor = concept; Security Kernel = implementation
7VM Escape is the most critical virtualisation attack
8SCADA priority = Safety first, NOT confidentiality (SAIC not CIA)
9AES block size ALWAYS 128 bits — key size varies
10ECB mode INSECURE — patterns leak in ciphertext
11DH = key exchange ONLY; DSA = signatures ONLY; RSA = everything
12HMAC cannot provide non-repudiation (shared secret key)
13Digital signatures = only way to get non-repudiation
14Birthday attack = 2^(n/2); 128-bit hash gives only 64-bit collision resistance
15RA verifies identity but does NOT sign certificates
16Root CA must be OFFLINE (air-gapped)
17NEVER escrow signing keys (destroys non-repudiation)
18Side-channel attacks exploit implementation, NOT algorithm math
19TLS 1.3 mandates PFS (only DHE/ECDHE allowed)
20Container isolation WEAKER than VM (shared kernel)
21Dedicated mode = simplest; Multilevel = most complex (different clearances coexist; needs MAC)
22Covert timing channels are HARDER to detect than storage channels
23Cold boot attack extracts keys from RAM shortly after power-off (freeze chips)
24Private cloud ≠ on-premises; private cloud can be hosted off-site (key = exclusive use)
25TCSEC B2 requires covert storage channel ANALYSIS; B3 adds covert timing channel analysis & MITIGATION
26Security depends on keeping algorithm secret (instead of key secrecy)
27Quantum computing breaks ALL encryption equally
28One-time pad is always unbreakable in practice
29Zero-knowledge proofs require revealing the secret
30Stream ciphers and block ciphers are interchangeable
31Confusion and diffusion are the same thing

Domain 4: Communication & Network Security

13% of Exam · Priority ★★★☆☆

4.1 OSI & TCP/IP Models

OSI #LayerTCP/IPProtocols / DevicesPDU
7ApplicationApplicationHTTP, HTTPS, FTP, SMTP, DNS, SNMP, LDAPData
6PresentationSSL/TLS encryption, JPEG, MPEG, ASCII, compressionData
5SessionRPC, NetBIOS, PPTP, SIP, NFSData
4TransportTransportTCP (reliable), UDP (fast), TLSSegment
3NetworkInternetIP, ICMP, IPSec, routers, L3 switchesPacket
2Data LinkNetwork AccessEthernet, ARP, switches, bridges, MAC addressesFrame
1PhysicalCables, hubs, repeaters, connectors, electrical signalsBits
🧠 "Please Do Not Throw Sausage Pizza Away" (L1→L7) / "All People Seem To Need Data Processing" (L7→L1)

Converged & Multilayer Protocols

ProtocolDescriptionSecurity Concern
VoIP (Voice over IP)Voice traffic over data networksEavesdropping (SRTP for encryption), vishing, toll fraud, QoS dependency
iSCSISCSI storage commands over TCP/IPStorage traffic on shared network; requires CHAP auth + IPSec or VLAN isolation
FCoE (Fibre Channel over Ethernet)Storage + network on same fabricShared infrastructure risk; requires dedicated VLANs
InfiniBandHigh-speed interconnect (HPC, data centres)Low-latency but limited built-in security; physical access control critical
Compute Express Link (CXL)High-speed CPU-to-device interconnect (memory, accelerators)Emerging; shared memory pools between hosts; data leakage between tenants; isolation enforcement critical
⚠ Multilayer protocol risk: Encapsulated protocols may bypass security controls that only inspect one layer (e.g., VPN within VPN, IPv6 tunnelled inside IPv4).

Network Performance Metrics

MetricDefinition
BandwidthMaximum data transfer rate (capacity of the link)
LatencyTime for packet to travel source → destination
JitterVariation in latency (critical for VoIP/video)
ThroughputActual data successfully transferred per unit time
SNR (Signal-to-Noise)Signal strength vs interference; higher = cleaner signal

Traffic Flow Directions

North-South: Traffic entering/leaving the network (client ↔ data centre); monitored by perimeter firewalls.
East-West: Traffic moving laterally within the network (server ↔ server); monitored by microsegmentation and internal IDS/IPS. East-west traffic often exceeds north-south in modern data centres.

Network Planes & Switching Architecture

ConceptDescriptionSecurity Note
Data PlaneForwards actual user traffic (packets/frames)Where DDoS and MitM occur
Control PlaneMakes routing/switching decisions (routing tables, ARP)Target of BGP hijacking, ARP poisoning; SDN centralises this
Management PlaneConfigures & monitors devices (SSH, SNMP, APIs)Most sensitive — out-of-band management preferred; restrict to jump server; MFA required
Cut-Through SwitchingForwards frame after reading destination MAC (low latency)No error checking — corrupted/malicious frames pass through
Store-and-ForwardReceives entire frame, checks CRC, then forwardsHigher latency but catches errors; preferred for security
⚠️ Out-of-band management uses a separate network (console, dedicated VLAN, serial) to manage devices — prevents attackers on the data plane from reaching management interfaces. In-band management (same network) is convenient but riskier.

Network Observability & Monitoring

Network Observability goes beyond traditional monitoring: combines logs, metrics, and traces to understand network behaviour holistically.
NetFlow / sFlow / IPFIX: Traffic flow metadata for analysis (who talks to whom, how much)
Capacity Management: Monitor bandwidth utilisation; plan for growth; prevent bottlenecks during DR failover
Fault Detection & Handling: Automated alerts on link failures, device unreachability; self-healing (FHRP, LACP failover)

Transmission Media

MediaTypeSecurity Characteristic
UTP/STPCopperUTP susceptible to EMI/eavesdropping; STP = shielded
CoaxialCopperMore resistant to EMI than UTP; legacy
Fibre OpticLightImmune to EMI; extremely difficult to tap; preferred for secure links

4.2 Network Devices & Segmentation

DeviceLayerFunction
HubL1Broadcasts all traffic to all ports (INSECURE)
SwitchL2Forwards by MAC address; creates collision domains
RouterL3Routes by IP; creates broadcast domains; ACLs
FirewallL3-L7Filters traffic by rules (stateless/stateful/application)
WAFL7Web Application Firewall; protects against XSS, SQLi, OWASP Top 10
IDS/IPSL3-L7Detect (IDS) / Prevent (IPS) intrusions
ProxyL7Forward proxy (client), Reverse proxy (server)

Firewall Types

TypeLayerHow It Works
Packet Filter (Stateless)L3Examines IP/port per packet; no connection tracking
Stateful InspectionL3-L4Tracks connection state (SYN/ACK); allows return traffic
Application ProxyL7Terminates & rebuilds connections; deep inspection
NGFW (Next-Gen)L3-L7Stateful + DPI + application awareness + IPS + threat intel
✓ Exam Tip: Stateful = tracks entire conversation. Stateless = each packet evaluated independently (faster but weaker).

Network Segmentation

ConceptPurpose
VLANLogical L2 segmentation; isolates broadcast domains
DMZBuffer zone between internet & internal network; hosts public-facing servers
Air GapPhysical isolation; no network connection (highest security)
MicrosegmentationFine-grained, workload-level segmentation (SDN/Zero Trust)
Jump Server / Bastion HostHardened access point for admin; single entry to secure zone

4.3 IP Addressing & Routing Protocols

IPv4 vs IPv6

FeatureIPv4IPv6
Address Size32-bit (4.3B addresses)128-bit (3.4×10³⁸ addresses)
FormatDotted decimal (192.168.1.1)Hex colon (2001:db8::1)
IPSecOptionalBuilt-in (mandatory support)
NATRequired (address scarcity)Not needed
BroadcastYesNo (uses multicast/anycast)
⚠ Exam Trap: IPv6 has IPSec built-in, but using it is NOT mandatory — implementation support is mandatory, not use.

Routing Protocols

ProtocolTypeAlgorithmScope
RIPDistance vectorHop count (max 15)Small networks
OSPFLink stateDijkstra (cost)Enterprise (IGP)
BGPPath vectorPolicy-basedInternet backbone (EGP)

4.4 Secure Protocols

InsecureSecure ReplacementWhat Changed
HTTP (80)HTTPS (443)TLS encryption
Telnet (23)SSH (22)Encrypted remote shell
FTP (21)SFTP (22) / FTPS (990)SSH tunnel / TLS
SNMP v1/v2 (161)SNMPv3Authentication + encryption
DNS (53)DNSSECDigital signatures on DNS records (integrity, NOT confidentiality)
LDAP (389)LDAPS (636)TLS encryption
POP3 (110) / IMAP (143)POP3S/IMAPS (995/993)TLS encryption
⚠ DNSSEC provides authentication/integrity but NOT confidentiality. DNS over HTTPS (DoH) or DNS over TLS (DoT) provide confidentiality.

TLS 1.3 Key Facts

TLS 1.3: 1-RTT handshake (vs. 2-RTT in 1.2) · Mandates PFS (ECDHE only) · Removed weak ciphers (RC4, 3DES, CBC) · No RSA key exchange (RSA still used for certificate authentication/signatures) · Only AEAD ciphers (AES-GCM, ChaCha20-Poly1305)
💡 See also: Sections 3.9–3.14 (D3) — Cryptography for the cipher suites, hashing algorithms, key exchange mechanisms (ECDHE/RSA/DH), and asymmetric foundations that underpin TLS, DNSSEC, S/MIME, and IPSec. Understanding the crypto layer explains why TLS 1.0/1.1 are deprecated.

4.5 VPN Technologies

IPSec

ComponentFunction
AH (Auth Header)Integrity + Authentication (NO encryption)
ESP (Encapsulating Security Payload)Confidentiality + Integrity + Authentication
IKEKey exchange & SA negotiation (UDP 500)
ModeProtectsUse Case
TransportPayload only (original IP header visible)Host-to-host
TunnelEntire packet (new IP header added)Gateway-to-gateway (site VPN)
🧠 "Tunnel Buries the Body": Tunnel mode = entire packet hidden. Transport = payload only.
⚠ AH provides NO encryption. If you need confidentiality, use ESP. AH is broken by NAT (NAT modifies the IP header, which invalidates AH's integrity check).

Other VPN Protocols

ProtocolLayerKey Feature
SSL/TLS VPNL4-L7Browser-based; no client needed; HTTPS (443)
L2TPL2No native encryption; pair with IPSec
WireGuardL3Modern, fast, minimal codebase

4.6 Wireless Security

StandardEncryptionAuthStatus
WEPRC4 (24-bit IV)Open / Shared Key❌ Broken (cracked in seconds)
WPATKIP (RC4 wrapper)PSK / 802.1X⚠️ Deprecated
WPA2AES-CCMPPSK / 802.1X (EAP)✅ Standard
WPA3AES-GCMP / SAESAE (Dragonfly) / 802.1X✅ Current best
WPA3-SAE: Simultaneous Authentication of Equals — replaces PSK; resistant to offline dictionary attacks; provides forward secrecy.
802.1X (EAP): Port-based NAC using RADIUS. Supplicant → Authenticator → Authentication Server.
⚠ WEP = instant fail. IV only 24-bit → exhausted in minutes → key recovery trivial.

Wireless Attacks

AttackDescription
Evil TwinRogue AP mimics legitimate SSID; MitM
DeauthenticationForces clients off network; capture WPA handshake
War DrivingScanning for open/weak wireless networks
Bluejacking/BluesnarfingBluetooth: unsolicited messages / data theft
KRACKKey Reinstallation Attack on WPA2 (patched)

4.7 Network Attacks

AttackLayerDescriptionMitigation
ARP PoisoningL2Fake ARP replies → MitMDAI, static ARP
MAC FloodingL2Overflow switch CAM table → hub modePort security
VLAN HoppingL2Double tagging / switch spoofingDisable DTP, prune unused VLANs
DNS PoisoningL7False DNS records in cacheDNSSEC
BGP HijackingL3Announce false routesRPKI, route filtering
SYN FloodL4Exhaust connection table (half-open)SYN cookies, rate limiting
Smurf AttackL3ICMP echo to broadcast with spoofed sourceDisable directed broadcast
MitM/MitBVariousIntercept/alter communicationsMutual TLS, certificate pinning

4.8 Email Security

ProtocolPurpose
SPFDNS record listing authorised mail servers for domain (IP-based)
DKIMCryptographic signature on email headers/body (integrity + authenticity)
DMARCPolicy layer atop SPF + DKIM; tells receivers what to do on failure (none/quarantine/reject)
S/MIMEPKI-based email encryption + signing (requires certificates)
PGP/GPGWeb-of-trust email encryption (no PKI hierarchy needed)
Email security stack: SPF + DKIM + DMARC together provide comprehensive anti-spoofing.

4.9 Modern Network Architectures

ConceptDescription
SDNSoftware-Defined Networking: separates control plane from data plane; centralised management
SD-WANSoftware-Defined WAN; optimises WAN connections; centralised policy
SASESecure Access Service Edge: SD-WAN + security (CASB, FWaaS, ZTNA, SWG) as cloud service
Zero Trust"Never trust, always verify." No implicit trust based on network location. Microsegmentation + continuous auth.
CDNContent Delivery Network: distributes content geographically; DDoS mitigation
VPC (Virtual Private Cloud)Logically isolated network within public cloud; own subnets, route tables, ACLs; equivalent of a private data centre in the cloud

Cellular & Mobile Networks

GenerationSecurity Note
4G/LTEIPSec between towers; mutual auth; but IMSI catchers (Stingray) can intercept
5GEnhanced encryption (256-bit), SUPI/SUCI privacy, network slicing (isolates virtual networks); expanded IoT attack surface

Secure Communication Channels (CBK 4.3)

ChannelSecurity Requirement
Voice / Video / CollaborationEnd-to-end encryption (SRTP for voice, TLS for signalling); meeting access controls; recording consent; screen-share DLP
Remote AccessVPN or ZTNA; MFA required; session timeout; administrative access via jump server/bastion host
Backhaul / SatelliteHigh latency; encryption mandatory (vulnerable to interception); consider link encryption for classified data
Third-Party ConnectivityDedicated links or VPN; SLA with security terms; monitoring of partner traffic; NAC enforcement
✓ Zero Trust Pillars: Verify explicitly · Use least-privilege access · Assume breach. Reference: NIST SP 800-207.

Domain 4 — Key Ports & Numeric Reference

Exam-Tested Port Numbers:
HTTP=80 · HTTPS=443 · SSH=22 · Telnet=23 · FTP control=21, data=20
SFTP=22 (SSH tunnel) · FTPS=990 (explicit 989) · DNS=53 · DHCP=67/68
SMTP=25 · SMTPS=465 · POP3=110 · POP3S=995 · IMAP=143 · IMAPS=993
LDAP=389 · LDAPS=636 · SNMP=161/162 · RDP=3389 · Kerberos=88
RADIUS=1812 auth / 1813 accounting · TACACS+=49 · BGP=179
IPSec/IKE=500 UDP · NAT-T (IPSec over NAT)=4500 UDP · L2TP=1701

Wireless Standards — Key Facts:
WEP: RC4 + 24-bit IV (exhausted in minutes — broken)
WPA: TKIP (RC4 wrapper) — deprecated
WPA2: AES-CCMP (128-bit key) — current standard
WPA3: AES-GCMP-256 + SAE (Dragonfly) — current best; resists offline dictionary attacks

TLS Version Status:
SSL 2.0 / 3.0 = broken (POODLE) · TLS 1.0 / 1.1 = deprecated
TLS 1.2 = acceptable (requires care with cipher suite selection)
TLS 1.3 = current standard: 1-RTT · PFS mandatory (DHE/ECDHE) · AEAD only · no RSA key exchange

Domain 4 — Mnemonics

#MnemonicHelps Remember
1"Please Do Not Throw Sausage Pizza Away"OSI layers L1→L7
2"Tunnel Buries the Body"IPSec Tunnel = entire packet; Transport = payload
3"AH = no confidentiality; ESP = everything"IPSec protocols
4"WEP = Wrecked Encryption Protocol"WEP is broken
5"SYN cookies solve SYN floods"DoS mitigation
6"SPF = who can Send, DKIM = Digital Key, DMARC = Decision Maker"Email security stack

Domain 4 — Exam Traps 🚨

#Trap
1IPv6 has mandatory IPSec SUPPORT, not mandatory USE
2DNSSEC = integrity/authentication only, NOT confidentiality (use DoH/DoT for that)
3AH provides NO encryption; broken by NAT (NAT modifies IP header, invalidating AH integrity check)
4TLS 1.3 = 1-RTT, mandates PFS, removed RSA key exchange
5WEP IV is only 24 bits → exhausted quickly → trivially cracked
6WPA3-SAE resists offline dictionary attacks (unlike WPA2-PSK)
7802.1X uses RADIUS (not TACACS+) for wireless enterprise auth
8Switches can still be compromised (MAC flooding, VLAN hopping)
9SSL VPN works through firewalls (port 443) — advantage over IPSec
10SDN separates control plane from data plane (centralised management)
11Zero Trust = "never trust, always verify" not "block everything"
12Smurf attack uses ICMP + broadcast + spoofed source (amplification)
13Management plane = most sensitive; must isolate via out-of-band management (separate VLAN/console)
14Cut-through switching is faster but passes corrupted frames; store-and-forward checks CRC (more secure)
15Data sovereignty = governed by country where data is STORED, not where company is headquartered
16Bluetooth attacks are all the same
17FTPS and SFTP are the same protocol
18L2TP provides encryption on its own
19SYN flood is hard to mitigate
20S/MIME and PGP use the same trust model
21Fraggle and Smurf are identical attacks
22DNS poisoning only impacts one target machine

Domain 5: Identity & Access Management (IAM)

13% of Exam · Priority ★★★☆☆

5.1 Identity Management Lifecycle

Identity lifecycle: Provisioning → Management → Review → Deprovisioning
Joiner-Mover-Leaver (JML): New hire = provision, Role change = modify, Termination = revoke (IMMEDIATELY)
ConceptDefinition
SubjectActive entity requesting access (user, process, device)
ObjectPassive entity being accessed (file, database, service)
PrincipalSubject after successful authentication

5.2 Authentication Factors

FactorTypeExamples
Type 1: Something You KnowKnowledgePassword, PIN, passphrase, security question
Type 2: Something You HavePossessionSmart card, token (TOTP/HOTP), phone, key fob
Type 3: Something You AreBiometricFingerprint, iris, retina, voice, face
Type 4: Somewhere You AreLocationGPS, IP geolocation, building access
Type 5: Something You DoBehavioralKeystroke dynamics, gait, signature
MFA: Must use 2+ DIFFERENT factor types. Two passwords = multi-step, NOT MFA.

Session Management

ControlPurposeKey Point
Session Timeout (Idle)Expire session after inactivity periodPrevents abandoned session hijacking; varies by sensitivity (15 min high-sec, 30 min normal)
Absolute TimeoutExpire session after max duration regardless of activityForces re-authentication; limits token validity window
Session ID ManagementUnique, random, unpredictable session tokensNever in URL (bookmarkable/logged); use secure cookies with HttpOnly + Secure + SameSite flags
Session Fixation PreventionGenerate NEW session ID after authenticationAttacker can’t pre-set a session ID and wait for victim to authenticate with it
Concurrent Session ControlLimit number of simultaneous sessions per userDetects credential sharing / compromise
⚠️ Exam loves session management: Session tokens must be regenerated after login. Never pass session IDs in URLs. Secure cookie attributes: HttpOnly (no JS access), Secure (HTTPS only), SameSite (CSRF protection).

Passwordless Authentication

Passwordless: Eliminates knowledge factor entirely. Uses possession + biometric instead.
Methods: FIDO2/WebAuthn (hardware authenticator + biometric), passkeys (synced across devices), magic links, push notification approval.
Benefits: Eliminates phishing of passwords, credential stuffing, password reuse. Aligns with Zero Trust identity verification.

Biometric Metrics

FRR (Type 1 Error): False Rejection Rate — rejects legitimate user
FAR (Type 2 Error): False Acceptance Rate — accepts impostor
CER/EER: Crossover Error Rate = where FRR = FAR (lower = more accurate)
Exam Rule: FAR is MORE DANGEROUS (lets intruder in)
⚠ Exam Trap: Type 1 error = False Rejection (FRR), NOT Type 1 factor. Type 2 error = False Acceptance (FAR), NOT Type 2 factor.

Retina vs Iris

BiometricScansKey Fact
RetinaBlood vessel pattern (back of eye)Most accurate; most invasive; reveals health conditions
IrisColoured ring patternVery accurate; less invasive; can use photo from distance

5.3 Single Sign-On (SSO) Technologies

TechnologyProtocolKey Feature
KerberosTicket-basedKDC (AS + TGS); uses symmetric crypto (PKINIT extension adds PKI for smart card/PIV authentication); TGT lasts ~10h; time-sensitive (5-min clock skew)
SAMLXML assertionsWeb SSO; IdP + SP; browser redirects; enterprise SSO
OAuth 2.0AuthorisationDelegated authorisation (NOT authentication); access tokens; "Login with Google"
OpenID Connect (OIDC)AuthenticationIdentity layer ON TOP of OAuth 2.0; ID tokens (JWT)
RADIUSAAAEncrypts password only; UDP 1812/1813; wireless, VPN, dial-up
TACACS+AAAEncrypts entire payload; TCP 49; granular command-level authz; Cisco

Kerberos Flow

1. User → AS: "I am Alice" (+ password hash)
2. AS → User: TGT (Ticket-Granting Ticket)
3. User → TGS: "I want access to FileServer" + TGT
4. TGS → User: Service Ticket
5. User → FileServer: Service Ticket → Access granted
⚠ Kerberos Attacks: Golden Ticket (forged TGT, compromised KRBTGT), Silver Ticket (forged service ticket), Pass-the-Ticket, Kerberoasting (offline cracking of service tickets).

RADIUS vs TACACS+

FeatureRADIUSTACACS+
TransportUDPTCP
EncryptionPassword onlyEntire payload
AAACombined A&ASeparate A/A/A
Best ForNetwork access (wireless, VPN)Device admin (routers, switches)
💡 See also: Section 3.14 Cryptographic Attacks (D3) for Kerberos-specific attacks — Pass-the-Hash, Golden Ticket (forged TGT), Silver Ticket (forged service ticket), and Kerberoasting. These are frequently tested Kerberos exam topics. Also see Section 5.3 for the SSO/Kerberos architecture that makes these attacks possible.

5.4 Access Control Models

ModelDecision ByDescription
DACData OWNEROwner sets permissions (ACLs); flexible; most common in desktops; risk = too permissive
MACSYSTEM / labelsSystem enforces based on classification labels; rigid; military/government; BLP/Biba models
RBACROLEAccess based on job role; groups/roles; enterprise standard; eases admin
ABACATTRIBUTESPolicies evaluate subject + object + environment attributes; most flexible; XACML
Risk-BasedRISK SCOREDynamic; adjusts access based on real-time risk assessment (user behaviour, location, device, threat intel); used in adaptive/step-up authentication
Rule-BasedRULESIF-THEN rules (firewalls, router ACLs); not same as RBAC

Access Policy Enforcement Architecture

PDP (Policy Decision Point): Evaluates access request against policies and makes allow/deny decision.
PEP (Policy Enforcement Point): Enforces the PDP's decision; actually blocks or permits access.
PAP (Policy Administration Point): Where policies are created and managed.
PIP (Policy Information Point): Provides attribute data (user roles, resource labels, environment) to the PDP.
XACML implements this architecture for ABAC systems.
⚠ RBAC ≠ Rule-Based. RBAC = Role-Based; Rule-Based = IF condition THEN action (e.g., firewall rules).

5.5 Privileged Access Management (PAM)

ControlPurpose
Just-In-Time (JIT)Elevated privileges granted only when needed, automatically revoked
Password VaultingPrivileged passwords checked out, auto-rotated, audited
Session RecordingRecord all admin sessions for audit & forensics
Least PrivilegeMinimum access for job function; no standing admin rights
Service AccountsNon-interactive; use managed identities; no interactive login

Service Account Management (Exam Pain Point)

Best PracticeRationale
No interactive loginService accounts should NEVER be used by humans to log in
Managed identitiesCloud-native identity (Azure MI, AWS IAM Roles) — no stored credentials to steal
Automated password rotationRotate credentials frequently via PAM vault; avoid hardcoded passwords
Least privilege scopingGrant only specific API permissions needed; never domain admin
Inventory & ownershipEvery service account must have a documented human owner; orphan service accounts = major risk
Monitoring & alertingAlert on interactive login attempts, unusual hours, privilege escalation from service accounts
⚠️ Service accounts are a top attack vector. Hardcoded credentials in scripts/repos, shared passwords, and orphan accounts (original creator left) are common audit findings. Treat them as privileged accounts.

Privilege Escalation Attacks

TypeDescriptionExample
VerticalLower-privilege user gains higher-privilege access (user → admin)Kernel exploit, sudo bypass
HorizontalSame-level user accesses another user's resourcesIDOR — accessing another user's account data via URL manipulation

Risk-Based / Adaptive Authentication

Adaptive authentication dynamically adjusts authentication requirements based on risk score. Factors include: user location, device, time, behaviour pattern, network. Higher risk → stronger authentication (step-up MFA).

💡 NIST SP 800-63 Identity Assurance Levels:
IAL 1: Self-asserted identity (no proofing) · IAL 2: Remote or in-person proofing (evidence verified) · IAL 3: In-person proofing + physical verification (highest confidence)
AAL 1/2/3 = Authenticator assurance levels (single/multi-factor/hardware crypto)

5.6 Access Review & Provisioning

User Access Reviews (UARs): Periodic review of user access rights; detect privilege creep; regulatory requirement.
ConceptDefinition
Privilege CreepAccumulation of unnecessary access over time (role changes without revocation)
Need-to-KnowEven with clearance, access only if job-function requires it
Separation of DutiesNo single person controls entire critical process; prevents fraud
Dual ControlTwo+ people must act together (nuclear launch = two keys)

Federated Identity

Federation: Trust relationship between organisations allowing users to access resources across domains using one identity.
IdP (Identity Provider): Authenticates users (e.g., Okta, Azure AD)
SP (Service Provider): Relies on IdP assertion; provides resource access

Domain 5 — Key Formulas & Numeric Facts

Biometric Error Rate Relationships:
FRR (Type 1 Error) = False Rejection Rate — legit user rejected (usability problem)
FAR (Type 2 Error) = False Acceptance Rate — impostor accepted (security risk — more dangerous)
CER / EER = Crossover Error Rate = the operating point where FRR = FAR
Lower CER = better overall biometric · Decreasing FAR increases FRR (tension between security and usability)

Kerberos Timing Constants:
Max clock skew tolerance: 5 minutes (authentication fails if clocks differ more)
Default TGT lifetime: ~10 hours · Service ticket lifetime: typically 5–8 hours
Golden Ticket attack: forges TGT using compromised KRBTGT hash — persists until KRBTGT is rotated twice

IAM Numeric References:
NIST SP 800-63 IAL 1: self-asserted · IAL 2: remote/in-person proofing · IAL 3: in-person + physical
NIST SP 800-63 AAL 1: single-factor · AAL 2: MFA required · AAL 3: hardware crypto authenticator
SAML assertion validity: typically 5 minutes (short window to prevent replay)

Domain 5 — Mnemonics

#MnemonicHelps Remember
1"I Am, I Prove, I Can"Identification → Authentication → Authorisation
2"Know Have Are"Authentication factor types 1-2-3
3"FAR is worse than FRR" (letting intruder IN)False Acceptance more dangerous
4"CER = sweet spot"Where FRR=FAR; lower CER = better biometric
5"Kerberos needs a clock"5-min clock skew tolerance
6"OAuth = Authorisation; OIDC = Authentication"OAuth alone is NOT authentication
7"RADIUS = Remote; TACACS+ = Terminal"RADIUS for network access; TACACS+ for device admin
8"DAC = Owner; MAC = System; RBAC = Role; ABAC = Attribute"Access control models

Domain 5 — Exam Traps 🚨

#Trap
1Two passwords = NOT MFA (same factor type); must be DIFFERENT factor types
2Type 1 error = False Rejection (FRR); Type 2 error = False Acceptance (FAR) — don't confuse with auth factor types
3OAuth 2.0 is AUTHORISATION only — use OIDC for authentication
4SAML uses XML; OIDC uses JWT tokens (modern, mobile-friendly)
5Kerberos = symmetric crypto + time-dependent; Golden Ticket = compromised KRBTGT
6RADIUS encrypts password only; TACACS+ encrypts entire payload
7RBAC ≠ Rule-Based Access Control (common confusion)
8Retina scan = most accurate BUT most invasive; can reveal health info
9Privilege creep = accumulation without revocation (fix with periodic UARs)
10DAC is most permissive (owner decides); MAC is most restrictive (system decides)
11Federation = cross-org trust; SSO = same organisation, multiple apps
12Smart card = possession (Type 2) + PIN = knowledge (Type 1) = MFA
13Risk-based access control is DYNAMIC (changes per session); RBAC is STATIC (changes per role assignment)
14Session tokens must be regenerated after login (prevent fixation); never in URLs (bookmarks/logs leak them)
15Service accounts = top attack vector; hardcoded creds in scripts, orphan accounts, no rotation = audit failures
16Managed identities (cloud-native) eliminate stored credentials entirely — preferred over service account passwords
17Identity proofing = authentication
18RBAC is the most granular access model
19Privileged accounts need standing access
20PAM and IAM are the same thing
21Password spraying = brute force
22Authentication only happens at login

Domain 6: Security Assessment & Testing

12% of Exam · Priority ★★★☆☆

6.1 Assessment Strategies

TypeDescriptionScope
Vulnerability AssessmentIdentifies known weaknesses (scanning tools)Breadth — "What's exposed?"
Penetration TestActively exploits vulnerabilitiesDepth — "Can we get in?"
Security AuditFormal evaluation against standard/policyCompliance — "Are rules followed?"
Red TeamAdversary simulation (stealth, multi-vector)Realistic attack — "How would APT do it?"
Blue TeamDefensive operations; detection & responseDefence — "Can we detect it?"
Purple TeamRed + Blue collaborate openlyImprovement — "Learn together"

6.2 Penetration Testing

Testing Knowledge Levels

TypePrior KnowledgeSimulates
Black Box (External)Zero knowledgeOutside attacker
White Box (Internal)Full knowledge (source, diagrams)Insider threat / thorough audit
Grey BoxPartial knowledgeCompromised user / partner

Penetration Test Phases

1. Planning/Scoping2. Reconnaissance (passive/active) → 3. Scanning/Enumeration4. Exploitation5. Post-Exploitation6. Reporting

Rules of Engagement (ROE)

✓ Written authorisation is MANDATORY before any penetration test. ROE defines: scope (in/out), timing, methods allowed, escalation contacts, emergency stop procedures, data handling.

6.3 Vulnerability Management

CVSS (Common Vulnerability Scoring System)

ScoreSeverity
0.0None
0.1–3.9Low
4.0–6.9Medium
7.0–8.9High
9.0–10.0Critical
CVE: Common Vulnerabilities & Exposures — unique identifier (CVE-YYYY-NNNNN)
CPE: Common Platform Enumeration — names products
NVD: National Vulnerability Database — NIST's CVE + CVSS enrichment

Scanning Types

Scanner TypeWhat It Finds
Network ScannerOpen ports, services, OS fingerprinting
Vulnerability ScannerKnown CVEs, misconfigurations, missing patches
DASTDynamic Application Security Testing — tests running app (black-box)
SASTStatic Application Security Testing — analyses source code (white-box)
IASTInteractive — combines SAST + DAST; agent inside running app
SCASoftware Composition Analysis — finds vulnerable libraries/dependencies

Additional Testing Techniques (CBK 6.2)

TechniqueDescription
Breach Attack Simulation (BAS)Automated, continuous simulation of real attack paths across kill chain; validates control effectiveness without manual red team
Compliance ChecksAutomated verification of systems against baseline standards (CIS, DISA STIG, PCI); often integrated into SIEM/config management

Test Output Analysis & Reporting (CBK 6.4)

ActivityDescription
RemediationFix identified vulnerabilities by priority (CVSS + business context); track to closure
Exception HandlingDocument accepted risks where remediation is not feasible; formal exception with compensating controls, expiry date, and management sign-off
Ethical DisclosureIf testing reveals vulnerabilities in third-party products: responsible disclosure to vendor → allow time to patch → coordinate public disclosure

6.4 Security Audits & Log Management

Audit Types

TypePerformed ByPurpose
Internal AuditInternal team (reports to board/audit committee)Self-assessment; identify issues early
External AuditIndependent third partyRegulatory compliance; SOC reports
Third-Party AuditCustomer/partner auditorsSupply chain assurance

SOC Reports

ReportScopeAudience
SOC 1Financial reporting controls (ICFR)Financial auditors
SOC 2Trust Services Criteria (Security, Availability, PI, Confidentiality, Privacy)Management, regulators, customers
SOC 2 Trust Services Criteria (TSC) — “SAPCP”:
Security (Common Criteria) — always tested; foundation for all other criteria
Availability — system uptime per SLA
Processing Integrity — system processing is complete, valid, accurate, timely, authorised
Confidentiality — information designated as confidential is protected
Privacy — personal information collected/used/retained/disclosed per privacy notice & GAPP
Security is MANDATORY in every SOC 2 report. Other criteria are selected based on the service’s commitments.
SOC 3Same as SOC 2 but general-use (summary)Public/marketing
TypePeriodDescription
Type IPoint-in-timeControls designed & implemented at a specific date
Type IIOver a period (6-12 months)Controls operating effectively over time (MORE VALUABLE)
✓ SOC 2 Type II is the gold standard for cloud/SaaS vendor assurance.
SSAE-18 (Statement on Standards for Attestation Engagements): Current auditing standard for SOC reports. Evolution: SAS 70 → SSAE-16 → SSAE-18. Requires management assertion and sub-service organisation monitoring.

Log Management

SIEM: Security Information & Event Management — collects, correlates, analyses logs; real-time alerts; dashboards.
SOAR: Security Orchestration, Automation & Response — automates incident response playbooks.
ConceptDefinition
Log IntegrityWrite-once storage, hash chains, centralised logging (prevents tampering)
NTPTime synchronisation critical for log correlation (all systems same clock)
RetentionLogs retained per policy/regulation; balance storage vs compliance

6.5 Software Testing Methods

Software Testing Levels (order of execution)

LevelScopeWho
Unit TestingIndividual functions/methodsDevelopers
Integration TestingCombined modules; data flows between componentsDevelopers / QA
System TestingComplete system end-to-endQA team
User Acceptance Testing (UAT)Business requirements met; real-world scenariosEnd users / business
Regression TestingVerify changes didn't break existing functionalityQA (after every change)

Testing Methods

MethodDescription
Code ReviewManual inspection of source code (most effective for logic flaws)
Fagan InspectionFormal 6-step code review: Planning → Overview → Preparation → Inspection → Rework → Follow-up (most rigorous)
Fuzz TestingRandom/malformed input to find crashes, buffer overflows, exceptions
Interface TestingTest APIs, UIs, data flows between components
Misuse/Abuse Case TestingTest from attacker's perspective; negative testing
Test Coverage AnalysisMeasure % of code exercised by tests (branch, statement, path, condition, loop)

Synthetic vs Real User Monitoring

TypeHowWhen
Synthetic MonitoringScripted transactions simulate user actions proactivelyPre-production; 24/7 baseline
Real User Monitoring (RUM)Captures actual user interactions passivelyProduction; real-world performance

6.6 Security Metrics & KPIs

KPIs vs KRIs (CBK 6.3 — Commonly Confused)

Metric TypeFull NamePurposeExample
KPIKey Performance IndicatorMeasures how WELL security controls and processes are performing% patches applied within SLA, mean time to detect, phishing click rate
KRIKey Risk IndicatorProvides EARLY WARNING of increasing risk exposure# unpatched critical CVEs, employee turnover in security team, % systems out of compliance
⚠ KPI = "How well are we doing?" (performance/effectiveness). KRI = "Are we heading toward trouble?" (risk trending). Both feed into management review and reporting. Don't confuse them — KPIs are backward-looking metrics of performance; KRIs are forward-looking indicators of potential risk.

Security Process Data Collection (CBK 6.3)

Data AreaWhat to Collect
Account ManagementOrphan accounts, privilege creep, access review completion rate
Management ReviewSecurity steering committee minutes, risk register updates, policy approval records
Backup VerificationBackup success/failure logs, restore test results (test restores regularly!), RPO verification
Training & AwarenessCompletion rates, phishing simulation results, security champion participation
DR/BCDRP test results, RTO/RPO achievement, plan update records

Operational Metrics

MetricMeasures
MTTDMean Time To Detect — how long before threat is discovered
MTTRMean Time To Respond/Remediate — how long to fix
MTTFMean Time To Failure — avg time until first failure (non-repairable)
MTBFMean Time Between Failures — avg time between failures (repairable)
Patch LatencyTime from patch release to deployment
Vulnerability DensityVulnerabilities per 1000 lines of code

Domain 6 — Key Formulas & CVSS Reference

CVSS v3.x Severity Scale:
0.0 = None · 0.1–3.9 = Low · 4.0–6.9 = Medium · 7.0–8.9 = High · 9.0–10.0 = Critical
CVSS base score = f(Attack Vector, Complexity, Privileges Required, Scope, CIA Impact)

Operational Reliability Metrics:
MTTD = Mean Time To Detect — elapsed time from intrusion/breach to discovery
MTTR = Mean Time To Respond/Remediate — time from detection to full remediation
MTTF = Mean Time To Failure — avg lifespan before first failure (non-repairable; no repair assumed)
MTBF = Mean Time Between Failures — avg time between failures (repairable systems; includes repair time)
Relationship: MTBF = MTTF + MTTR for repairable systems

SOC Report Types — Quick Reference:
SOC 1: financial reporting controls (ICFR) · Audience: financial auditors
SOC 2: Trust Services Criteria (Security mandatory + optional: Availability, PI, Confidentiality, Privacy)
SOC 3: SOC 2 summary — public marketing document
Type I = at a point in time · Type II = over 6–12 months (strongest assurance)

Domain 6 — Mnemonics

#MnemonicHelps Remember
1"VA = breadth; Pentest = depth"Vulnerability assessment vs penetration test
2"Black = blind; White = wide open; Grey = glimpse"Pentest knowledge levels
3"SOC 1 = $; SOC 2 = Security; SOC 3 = Summary"SOC report types
4"Type I = snapshot; Type II = movie"Point-in-time vs over time
5"DAST = Dynamic = running app; SAST = Static = source code"App testing types

Domain 6 — Exam Traps 🚨

#Trap
1Vulnerability scan ≠ penetration test (scan identifies; pentest exploits)
2Written authorisation REQUIRED before any pentest (no verbal OK)
3SOC 2 Type II > Type I for ongoing assurance
4Fuzz testing finds bugs code review misses (crashes, memory corruption)
5MTBF includes repair time; MTTF = until first failure (no repair assumed)
6Log integrity requires centralised logging + tamper protection + NTP
7Red team = stealth adversary sim; Purple team = collaborative improvement
8SCA finds vulnerable open-source libraries (not your code, their bugs)
9KPI = performance (backward-looking); KRI = risk warning (forward-looking) — don't confuse them
10Backup verification = actually TEST restores regularly, not just check backup logs succeeded
11SOC 1 = SOC 2 = SOC 3
12100% code coverage = no bugs
13Patches don't need regression testing
14False positives are worse than false negatives
15Testing only covers what SHOULD work (not abuse cases)

Domain 7: Security Operations

13% of Exam · Priority ★★★☆☆

7.1 Incident Response (IR)

NIST SP 800-61 — IR Lifecycle

1. Preparation2. Detection & Analysis3. Containment, Eradication & Recovery4. Post-Incident Activity (Lessons Learned)
PhaseKey Actions
PreparationIR plan, team (CSIRT), tools, training, communication plan, legal counsel
Detection & AnalysisMonitoring (SIEM), IoC identification, triage, severity classification
ContainmentShort-term (isolate host) → Long-term (patch, rebuild) → Evidence preservation
EradicationRemove malware, close vulnerabilities, reset credentials
RecoveryRestore systems, verify functionality, enhanced monitoring
Lessons LearnedPost-mortem within 1-2 weeks; update IR plan; root cause analysis

DR Communications Plan (CBK 7.11 — Heavily Tested)

ElementDetail
Notification RosterCall tree / cascade: who calls whom, in what order; include alternates for each role
Stakeholder CommunicationExecutives, employees, customers, partners, regulators — each needs tailored messaging
Communication MethodsPrimary + backup channels (if email is down, use SMS/mass notification system/satellite phone)
SpokespersonDesignated media contact; ONLY authorised person speaks to press (usually PR/legal, NOT IT)
Regulatory NotificationsMandatory breach reporting (GDPR 72h, HIPAA 60d); regulator-specific requirements
Status UpdatesRegular cadence during DR activation; include recovery progress vs RTO targets
All-Clear DeclarationFormal notification that operations have been restored and are verified; authorised by management
⚠️ DR Communications is a common exam question. The plan must address: WHO communicates, WHAT they say, HOW (through which channels), WHEN (at what intervals), and to WHOM. Backup communication methods are essential — the disaster may have destroyed primary channels.
⚠ First Priority in IR: Protect HUMAN SAFETY (life safety always first), then Contain the incident (limit damage), then Preserve evidence.
💡 See also: Section 1.6 Investigation Types (D1) for legal standards of evidence (beyond reasonable doubt vs. preponderance), chain of custody requirements, and the difference between criminal, civil, administrative, and regulatory investigations — all of which must be considered from the first moment of Incident Response. Also see Section 7.2 Digital Forensics for evidence collection procedures during active IR.

7.2 Digital Forensics

Order of Volatility (most → least volatile)

1. Registers/Cache2. RAM3. Swap/Pagefile4. Running Processes/Network5. Disk6. Remote Logs7. Backup Tapes/Archives
🧠 "Real Rascals Swap Places, Detectives Lag Behind" — Registers, RAM, Swap/Pagefile, Processes, Disk, Logs, Backups

Evidence Handling

ConceptDefinition
Chain of CustodyDocumented history of evidence: who handled, when, what was done. UNBROKEN chain or evidence is inadmissible.
Forensic ImageBit-for-bit exact copy of original media (dd, FTK Imager). Hash before & after to prove integrity.
Write BlockerHardware/software preventing writes to evidence drive during imaging
Legal HoldPreserve all potentially relevant evidence; overrides retention policies

Evidence Types (strongest → weakest)

TypeDescriptionStrength
Real/PhysicalTangible objects (hard drive, printed document)Strongest
DocumentaryWritten records (contracts, logs)Strong (if authenticated)
TestimonialWitness statements (expert/fact)Moderate
DemonstrativeCharts, models, re-enactmentsSupporting
HearsaySecond-hand statementsGenerally inadmissible
✓ Evidence must be: Relevant (relates to case), Reliable (trustworthy collection), Sufficient (enough to prove), Legally permissible (properly obtained).

Cloud Forensics Challenges

ChallengeDescription
Jurisdictional IssuesData may span multiple countries; which laws apply? Provider’s HQ, data location, or victim’s?
Multi-TenancyShared infrastructure makes isolating evidence difficult; provider must cooperate
Volatile EvidenceVMs/containers can be destroyed instantly; auto-scaling removes instances; snapshots essential
Limited AccessCustomer may not have access to hypervisor, network logs, or physical media — depends on service model (IaaS > PaaS > SaaS access)
Chain of CustodyProvider must assist with forensic imaging; contractual agreements for forensic support needed BEFORE incident
Log AvailabilityCloud logs may have limited retention; must configure extended retention and centralised export proactively
💡 Cloud forensics tip: Negotiate forensic investigation rights in the cloud contract BEFORE an incident. SaaS gives you the LEAST forensic access; IaaS gives the MOST.

Insider Threat Indicators

CategoryIndicators
BehaviouralWorking unusual hours, excessive access requests, bypassing controls, hostility toward employer, financial stress
TechnicalBulk data downloads, USB use, accessing files outside scope, email to personal accounts, VPN at unusual times
Policy ViolationsDisabling security tools, sharing credentials, ignoring classification labels
Detection: UEBA (User & Entity Behaviour Analytics) + DLP + privileged access monitoring + mandatory vacation + job rotation. UEBA baselines normal behaviour and alerts on anomalies — most effective technical control for insider threats.

7.3 Disaster Recovery & High Availability

Recovery Sites

SiteEquipmentDataRecovery TimeCost
Hot Site✅ Active✅ Real-time/nearMinutes–hours$$$$
Warm Site✅ Partial⚠️ Needs restoreHours–days$$$
Cold Site❌ Empty❌ From backupDays–weeks$
Cloud/DRaaS✅ On-demand✅ ReplicatedMinutesVariable
Mobile Site✅ Portable⚠️Hours$$
ReciprocalPartner orgVariable$

RAID Levels

RAIDDescriptionMin DisksFault Tolerance
RAID 0Striping (performance, NO redundancy)2NONE
RAID 1Mirroring (exact copy)21 disk
RAID 5Striping + distributed parity31 disk
RAID 6Striping + double parity42 disks
RAID 10Mirror + Stripe (RAID 1+0)41 per mirror pair
⚠ RAID 0 = ZERO redundancy. RAID is NOT a backup — it's availability (protects against disk failure, not data corruption/deletion).

Backup Types

TypeWhat It Backs UpArchive BitRestore Needs
FullEverythingClears1 tape
IncrementalChanged since LAST backupClearsFull + ALL incrementals
DifferentialChanged since last FULLDoes NOT clearFull + LAST differential
🧠 "Incremental = each increment; Differential = difference from full"

Media Management & Protection (CBK 7.3)

ControlDescription
LabellingMark classification level on all removable media (tapes, USB, optical)
HandlingFollow procedures per classification; clean desk for media; log check-in/out
StorageLocked containers rated for classification; fire-rated safes for backups; off-site rotation
TransportEncrypt before transport; tamper-evident bags; bonded courier for classified media
SanitisationClearing (overwrite) → Purging (degauss/crypto-erase) → Destruction (shred/incinerate) per NIST 800-88
RetentionDefined retention schedule per policy/regulation; destroy when no longer needed

Quality of Service (QoS)

ConceptDescription
QoSPrioritise critical traffic (VoIP, video, SCADA) over best-effort traffic; essential during DR
DiffServDifferentiated Services — marks packets with DSCP values for priority routing
IntServIntegrated Services — reserves bandwidth end-to-end via RSVP
Traffic ShapingSmooth traffic bursts; delay non-critical packets to ensure SLA for priority flows
💡 QoS is tested in recovery/continuity context: During failover, bandwidth is limited — QoS ensures critical services (emergency comms, replication) are prioritised.
💡 See also: Section 1.11 Business Continuity Planning (D1) for the MTD, RTO, RPO, and WRT formulas that determine which recovery site tier is required — Hot (RTO < hours), Warm (RTO < days), Cold (RTO < weeks). The BIA produces these targets; DR planning must satisfy them.

7.4 Change & Configuration Management

Change Management Process: Request → Assess (impact, risk) → Approve (CAB) → Test → Implement → Review
Change TypeDescriptionApproval
StandardPre-authorised, low-risk, well-documented (e.g., password reset)Pre-approved — no CAB needed
NormalRoutine change with potential impact; follows full processCAB review & approval required
EmergencyUrgent fix for critical issue (e.g., zero-day patch)Expedited approval; retrospective CAB review
ConceptDefinition
Configuration BaselineKnown-good approved configuration state
Configuration Item (CI)Any component under config management (server, router, app)
CMDBConfiguration Management Database — tracks all CIs & relationships
CABChange Advisory Board — reviews/approves changes

Configuration Management Automation

Tool / ConceptDescription
Infrastructure as Code (IaC)Terraform, CloudFormation — define infrastructure in version-controlled templates; ensures consistent, repeatable deployments
Configuration Management ToolsAnsible, Puppet, Chef, SaltStack — enforce desired state; auto-remediate drift from baseline
Immutable InfrastructureNever patch running systems; replace with new pre-hardened images. Prevents configuration drift & unpatched states
Baseline Compliance ScanningCIS Benchmarks, DISA STIGs — automated scanning verifies systems match approved baseline; feed results to SIEM
💡 Automation is key to CBK 7.3. Manual config management doesn’t scale. “Provisioning, baselining, automation” are the three pillars the exam outline calls out explicitly.

7.5 Physical Security

Control LayerExamples
DeterFences, lighting, signs, security guards
DetectCCTV, motion sensors, intrusion alarms, dogs
DelayMultiple barriers, mantraps/vestibules, bollards
Deny/PreventLocks, access cards, biometrics, guards
Mantrap / Vestibule: Two-door entry; prevents tailgating/piggybacking. Only one door opens at a time.

CPTED in Operations

CPTED (Crime Prevention Through Environmental Design): Natural surveillance (sight lines) · Natural access control (single entry) · Territorial reinforcement (signs, maintained boundaries).

Fence Heights

HeightPurpose
3–4 ftBoundary marker; deters casual trespass
6–7 ftHard to climb; deters most intruders
8 ft + razor wireCritical/high-security areas

Fire Classes

ClassFuelSuppression
AOrdinary combustibles (wood, paper)Water, soda acid
BFlammable liquids (gas, oil)CO₂, FM-200, dry chemical
CElectrical equipmentCO₂, FM-200 (non-conductive)
DCombustible metals (magnesium)Dry powder (special agent)
KKitchen (cooking oils/fats)Wet chemical

Fire Suppression

AgentBest ForKey Concern
Water (wet/dry pipe)General officeElectronics damage; dry pipe for freezing
FM-200 / Novec 1230Data centresClean agent; safe for electronics & humans
CO₂Unmanned areasDisplaces oxygen → LETHAL to humans
HalonLegacyBanned (ozone depletion); FM-200 is replacement
⚠ CO₂ suppression in occupied areas = DEATH RISK. FM-200/Novec 1230 = safe for people.

Personnel Safety (CBK 7.15)

ConcernControl
DuressSilent alarm / duress code (e.g., reverse PIN, panic button)
Travel SecurityEncrypted laptops, VPN, no sensitive data on portable devices, tamper-evident bags
Emergency EvacuationDrills, assembly points, head counts, AEDs, first-aid kits
Workplace ViolenceVisitor logs, access controls, security guards, training
✓ FIRST priority in ANY emergency = protect human life. Always.

7.5b Malware Types & Detective/Preventative Measures

Malware TypeCharacteristicKey Fact
VirusRequires host program; self-replicatingNeeds user action to spread
WormSelf-propagating; no host neededSpreads via network autonomously
TrojanDisguised as legitimate softwareDoes not self-replicate
RansomwareEncrypts data; demands paymentPrimary target = Availability
RootkitHides deep in OS/firmwareKernel rootkit = reinstall OS; firmware rootkit = replace hardware
Logic BombTriggers on condition (date, event)Often planted by insider
SpywareCovert data collectionKeyloggers, screen capture
RAT (Remote Access Trojan)Backdoor remote controlFull system control for attacker
Fileless MalwareLives in memory; no files on diskEvades traditional AV; uses PowerShell/WMI
PolymorphicChanges code/signature each timeEvades signature-based detection
MetamorphicRewrites entire codeEven harder to detect than polymorphic
Anti-Malware Strategies:
Signature-based: Known patterns; fast but misses zero-day
Heuristic/Behavioural: Detects suspicious actions; higher false positives
Sandboxing: Isolate & execute suspicious files safely
EDR (Endpoint Detection & Response): Continuous monitoring + automated response
Allowlisting: Only approved executables run (strongest, most restrictive)
UEBA (User & Entity Behavior Analytics): Machine learning to detect anomalous user/device behaviour; complements SIEM

MFA Fatigue / Push Bombing

AspectDetail
AttackAttacker has stolen credentials; sends repeated MFA push notifications until user approves out of frustration
MitigationNumber matching (user must enter number shown on login screen), rate limiting push requests, phishing-resistant MFA (FIDO2)
⚠️ Push-based MFA alone is NOT phishing-resistant. FIDO2/WebAuthn hardware keys are the strongest defence against MFA fatigue and phishing.

Honeypots & Honeynets

TypeDescriptionLegal Note
HoneypotSingle decoy system designed to attract attackersEnticement = legal
HoneynetNetwork of honeypots; simulates entire infrastructureMore realistic; captures lateral movement
⚠️ Enticement (honeypot) = legal. Entrapment (forcing someone to commit crime) = illegal.

7.6 Investigation Types & eDiscovery

TypeStandardOutcome
CriminalBeyond reasonable doubtProsecution, jail, fines
CivilPreponderance of evidence (more likely than not)Monetary damages
AdministrativeLeast formal; organisational policyDiscipline, termination
RegulatoryGovernment agency investigationSanctions, fines, licence revocation

Electronic Discovery (eDiscovery)

eDiscovery is the process of identifying, collecting, and producing electronically stored information (ESI) in response to legal proceedings.

PhaseActivity
IdentificationLocate potentially relevant ESI across all systems
PreservationLegal hold — prevent destruction of relevant data; overrides retention policies
CollectionGather ESI in forensically sound manner
ProcessingReduce volume; de-duplicate; convert formats
ReviewAttorneys assess for relevance and privilege
ProductionDeliver to opposing counsel in agreed format
⚠️ Litigation Hold overrides ALL retention/destruction schedules. Failure to preserve = spoliation = severe legal sanctions.

7.7 Patch, Vulnerability & Threat Management

Patch Management Cycle: Identify → Evaluate → Test → Deploy → Verify → Document

Ingress & Egress Monitoring

DirectionMonitorsPurpose
IngressInbound traffic entering the networkBlock malicious payloads, scans, unauthorized access
EgressOutbound traffic leaving the networkDetect data exfiltration, C2 callbacks, policy violations
💡 Egress filtering is often neglected but critical. DLP, firewall rules, and proxy inspection prevent data exfiltration.

NIST SP 800-137: Continuous Monitoring

Establishes an Information Security Continuous Monitoring (ISCM) strategy: ongoing awareness of security posture, vulnerabilities, and threats. Integrates with RMF step 7 (Monitor). Uses automated tools (SIEM, vulnerability scanners, configuration checkers) for near-real-time visibility.

Restoration vs Recovery

TermDefinitionFocus
RecoveryBring IT operations back to functional state at alternate siteGetting systems running (DR site)
RestorationReturn operations to the original (or new permanent) primary facilityMoving back to normal operations
⚠️ Recovery = move to DR site. Restoration = move back to primary. During restoration, the DR site becomes the backup in case the primary fails again.
ConceptDefinition
Threat IntelligenceIoCs, TTPs, threat feeds; actionable info about adversaries
Threat HuntingProactive search for threats that evade automated detection
MITRE ATT&CKKnowledge base of adversary TTPs; maps real-world attack techniques

7.8 AI/ML Operational Threats

ThreatDescriptionMitigation
Model DriftModel accuracy degrades over time as real-world data distribution changes (data drift) or relationships shift (concept drift)Continuous monitoring, periodic retraining, performance thresholds with automated alerts
Data PoisoningAttacker manipulates training data to produce biased/wrong outputsInput validation, provenance tracking, anomaly detection on training data
Model InversionAttacker queries model to reconstruct sensitive training dataDifferential privacy, rate limiting queries, access controls on model APIs
Adversarial InputsCrafted inputs that cause model misclassification (e.g., fooling image recognition)Adversarial training, input preprocessing, ensemble models
💡 ISC2 added AI security across all domains in 2024. Know model drift, data poisoning, and adversarial ML — these are new exam topics.

Domain 7 — Key Formulas & Numeric Facts

Availability Formula:
Availability % = MTBF ÷ (MTBF + MTTR) × 100

High Availability “Nines” Reference:
99% = 3.65 days/yr downtime · 99.9% (3-nines) = 8.77 hrs/yr
99.99% (4-nines) = 52.6 min/yr · 99.999% (5-nines) = 5.26 min/yr

RAID Usable Capacity:
RAID 0: n × disk (no redundancy) · RAID 1: 1 disk (50% for 2 disks)
RAID 5: (n−1) × disk (1 parity disk lost) · RAID 6: (n−2) × disk (2 parity disks)
RAID 10: (n/2) × disk (50% overhead for mirror pairs)

Backup Restore Complexity:
Full only: 1 media set (fastest restore) · Full + Differential: 2 sets (Full + most recent Diff)
Full + Incremental: 1 + n sets (Full + ALL incrementals since last full — slowest restore)

RTO + WRT ≤ MTD (from D1 — applies equally to D7 DR operations)

Mandatory Breach Notification Timelines:
GDPR → Supervisory Authority: 72 hours from controller awareness
HIPAA → HHS (+ individuals if >500): 60 days from discovery
PCI-DSS → Card brands + acquirer: immediately / as soon as practical (contractual SLA)

Domain 7 — Mnemonics

#MnemonicHelps Remember
1"Prepare Detect Contain Learn"NIST IR phases
2"Real Rascals Swap Places, Detectives"Order of volatility (Registers, RAM, Swap, Processes, Disk…)
3"RAID 0 = ZERO protection"RAID 0 has no redundancy
4"Hot = Hours; Warm = Wait; Cold = Construct"Recovery site readiness
5"Incremental = each step; Differential = from Full"Backup types restore complexity
6"CO₂ = Coffin; FM-200 = Friendly"Fire suppression safety
7"Virus needs a ride; Worm drives itself"Virus = host + user; Worm = self-propagating
8"A-B-C-D-K" fire classesAsh (ordinary), Boil (liquids/gases), Current (electrical), Dynamite (combustible metals), Kitchen (cooking oils)
9"CPTED = See-Access-Territory"Natural surveillance, access control, territorial
10"Duress = Silent SOS"Covert distress signal, reverse PIN, panic button

Domain 7 — Exam Traps 🚨

#Trap
1FIRST priority in incident = human safety, THEN containment, THEN evidence
2Evidence without chain of custody = INADMISSIBLE
3Collect most volatile evidence FIRST (RAM before disk)
4Forensic images must be hashed before AND after (integrity proof)
5RAID is NOT backup — protects against disk failure only, not data loss
6RAID 0 = performance only; 1 disk fails = ALL data lost
7Incremental restore needs Full + ALL incrementals (slow restore)
8Differential restore needs Full + LAST differential only (faster restore)
9CO₂ fire suppression = lethal to humans; FM-200 = safe for people
10Halon is banned (ozone); FM-200 is the replacement
11Criminal = beyond reasonable doubt; Civil = preponderance of evidence
12Change management = formal CAB approval BEFORE implementation
13Mantrap prevents tailgating; CCTV detects but doesn't prevent
14Lessons learned meeting ideally within 1-2 weeks of incident
15Virus needs host + user action; Worm self-propagates across network autonomously
16Rootkit in firmware = replace hardware; OS rootkit = full reinstall
17Polymorphic changes signature; metamorphic rewrites entire code — both evade signature AV
18Allowlisting > blocklisting (default-deny is stronger than default-allow)
19Entrapment = illegal (inducing crime); enticement = legal (honeypot lures existing intent)
20Fire Class C = electrical — NEVER use water (electrocution risk)
21DR Communications plan must have BACKUP channels (primary comms may be destroyed by the disaster)
22Cloud forensics: SaaS = least forensic access; IaaS = most. Negotiate investigation rights in contract BEFORE incident
23UEBA = best technical control for insider threats (baselines normal behaviour, detects anomalies)
24IaC + config automation (Ansible/Puppet) enforces baseline compliance at scale — manual doesn’t work for large environments
25Recovery = move TO DR site; Restoration = move BACK to primary. During restoration, DR site becomes the backup
26Copies of evidence are always as admissible as originals
27Computer logs are inadmissible hearsay in all cases
28All evidence types have equal legal weight
29Criminal burden of proof applies to all investigations
30IDS catches everything
31SPAN port and network tap are equally reliable
32Clipping levels are unnecessary overhead
33Warm site = hot site without data

Domain 8: Software Development Security

10% of Exam · Priority ★★☆☆☆

8.1 SDLC Models

ModelApproachKey Feature
WaterfallSequential, linearNo going back; requirements fixed upfront; documentation-heavy
V-ModelWaterfall + testing at each stageVerification & validation mapped to each dev phase
SpiralIterative with risk analysisEach loop = risk assessment; prototyping; cost estimation
AgileIterative, incrementalSprints (2-4 weeks); working software; customer collaboration
ScrumAgile frameworkProduct Owner, Scrum Master, Sprint, Daily Standup, Sprint Review
DevOpsDev + Operations collaborationCI/CD pipelines; automation; rapid deployment
DevSecOpsDevOps + Security"Shift left" — integrate security from the start (SAST/DAST in pipeline)
Scaled Agile Framework (SAFe)Agile at enterprise scaleMulti-team coordination; Program Increments (PIs); Agile Release Trains (ARTs); adds governance layer to Agile for large organisations
"Shift Left": Move security testing earlier in SDLC — code commit triggers SAST, build triggers DAST; cheaper to fix early.

Integrated Product Team (IPT)

IPT: Cross-functional team (developers, security, QA, operations, business) that collaborates throughout the SDLC to ensure security is integrated from inception — not bolted on at the end.
RoleContribution
Security ArchitectThreat models, security requirements, design review
DeveloperSecure coding, fix vulnerabilities from SAST/DAST
QA / TestSecurity test cases, regression testing, fuzz testing
OperationsDeployment hardening, monitoring, incident response readiness
Product OwnerPrioritises security stories alongside features

COTS & Acquired Software Security (CBK 8.4)

SourceSecurity Considerations
COTS (Commercial Off-The-Shelf)Evaluate vendor security posture; review CVEs; ensure patch support; contractual SLAs for vulnerability response
Open SourceReview licence, community activity, known vulns (SCA); verify integrity of downloads (checksums/signatures)
Third-Party LibrariesSBOM (Software Bill of Materials); track dependencies; automated SCA in CI/CD pipeline
Managed / Cloud ServicesShared responsibility model; vendor SOC 2 reports; assess API security; data residency; exit strategy
💡 SBOM = inventory of all components in software. Required by US Executive Order 14028 for federal software. Enables rapid vulnerability identification (e.g., Log4Shell).

Software-Defined Security

Software-Defined Security: Security policies programmatically defined, deployed, and enforced via APIs and automation — decoupled from hardware.
ConceptDescription
Policy as CodeSecurity policies written in code (OPA/Rego, Sentinel); version-controlled, testable, auditable
Infrastructure as Code (IaC)Terraform, CloudFormation — security configs embedded in templates; drift detection
Security Orchestration (SOAR)Automated incident response playbooks; enrichment → decision → action

8.1b Object-Oriented Security Concepts

ConceptDefinitionSecurity Relevance
EncapsulationHides internal state; exposes only public interfaceData hiding prevents direct manipulation — enforces controlled access to object data
InheritanceChild class inherits properties/methods from parentInherited permissions can create unintended access; must review inherited security attributes
PolymorphismSame interface, different behaviour depending on object typeMalicious class could override methods — validate object type at runtime
AbstractionSimplifies complex systems; shows only relevant detailReduces attack surface by hiding implementation complexity
CouplingDegree of interdependence between modulesLoose coupling preferred — vulnerability in one module doesn't cascade to others
CohesionDegree to which module elements belong togetherHigh cohesion preferred — single-purpose modules are easier to secure and test
⚠ Exam Trap: Encapsulation = BEST security feature of OOP (data hiding). Inheritance can be a security risk (child inherits parent's vulnerabilities/permissions). Loose coupling + High cohesion = secure design.

8.2 Secure Coding & OWASP Top 10

OWASP Top 10 (2021)

#VulnerabilityMitigation
A01Broken Access ControlDeny by default, enforce server-side, least privilege
A02Cryptographic FailuresStrong algorithms, key management, encrypt at rest/transit
A03Injection (SQLi, XSS, LDAP, OS command)Parameterised queries, input validation, escaping
A04Insecure DesignThreat modelling, secure design patterns, abuse cases
A05Security MisconfigurationHardening, remove defaults, minimal install
A06Vulnerable ComponentsSCA scanning, patch management, SBOM
A07Authentication FailuresMFA, strong passwords, session management
A08Software & Data IntegrityVerify updates, code signing, CI/CD integrity
A09Logging & Monitoring FailuresCentralised logging, alert on suspicious events
A10SSRF (Server-Side Request Forgery)Validate URLs, whitelist destinations, network segmentation

Key Attack Types

AttackDescriptionPrevention
SQL InjectionMalicious SQL in input fieldsParameterised queries / prepared statements
XSS (Cross-Site Scripting)Inject malicious script into web pagesOutput encoding, CSP headers
CSRF (Cross-Site Request Forgery)Trick authenticated user into unintended actionAnti-CSRF tokens, SameSite cookies
Buffer OverflowWrite beyond allocated memoryBounds checking, ASLR, DEP/NX, safe languages
Race Condition (TOCTOU)Time-of-check vs time-of-use gap exploitedLocks, atomic operations
SSRFServer tricked into fetching internal resourcesWhitelist, disable unnecessary protocols

API Security (CBK 8.5)

ControlDescription
AuthenticationOAuth 2.0 / OpenID Connect for API auth; API keys for identification (NOT authentication alone)
AuthorisationScope-based access; enforce least privilege per endpoint; RBAC on API resources
Rate LimitingThrottle requests to prevent abuse/DoS; implement per-user and per-IP limits
Input ValidationValidate all parameters, headers, body; reject unexpected fields; type checking
API GatewayCentralised enforcement point for auth, rate limiting, logging, TLS termination
VersioningDeprecate old API versions gracefully; avoid breaking changes; sunset policy
DocumentationOpenAPI/Swagger specs; avoid exposing internal endpoints; restrict discovery in production
⚠️ OWASP API Security Top 10 is separate from Web Top 10. Key API risks: Broken Object Level Auth (BOLA), mass assignment, excessive data exposure.

Secure Coding Principles

Input Validation: Whitelist (allow known-good) > Blacklist (block known-bad)
Output Encoding: Encode data before rendering (prevents XSS)
Least Privilege: App runs with minimum needed permissions
Defence in Depth: Multiple layers of controls
Fail Secure: On error, deny access (not fail-open)

8.3 Database Security

ConceptDefinition
AggregationCombining low-classification data to derive higher-classification information
InferenceDeducing sensitive info from non-sensitive data + metadata
PolyinstantiationMultiple rows with same key but different classification levels (prevents inference)
ViewsVirtual table limiting what users see (constrained interface)
Stored ProceduresPrecompiled DB operations; prevent direct SQL access

ACID Properties (Transactions)

Atomicity — all or nothing (commit or rollback)
Consistency — DB stays in valid state
Isolation — concurrent transactions don't interfere
Durability — committed data survives crashes

Database Types

TypeDescription
Relational (SQL)Tables, rows, columns; structured; ACID; SQL (MySQL, PostgreSQL, Oracle)
NoSQLDocument, key-value, graph, columnar; flexible schema; BASE (MongoDB, Cassandra)
Data WarehouseHistorical aggregated data; OLAP; decision support
Data LakeRaw unstructured/structured data; schema-on-read; big data analytics

Database Normalization (Classic Exam Topic)

Normal FormRequirementEliminates
1NFEach cell contains a single atomic value; no repeating groupsDuplicate data in rows
2NF1NF + all non-key attributes fully depend on the ENTIRE primary keyPartial dependencies (attributes depending on part of composite key)
3NF2NF + no transitive dependencies (non-key attributes don't depend on other non-key attributes)Transitive dependencies
Why Normalization Matters for Security: Reduces data redundancy → fewer places to protect → referential integrity maintained → prevents update/insertion/deletion anomalies. Denormalization (intentionally adding redundancy) improves read performance but increases data inconsistency risk.
🧠 "The Key, the Whole Key, and Nothing but the Key (so help me Codd)" — 1NF = the key; 2NF = the whole key; 3NF = nothing but the key

Software Escrow

Software Escrow: Source code and documentation deposited with a neutral third-party escrow agent. Released to the customer ONLY if specific trigger conditions are met (vendor bankruptcy, failure to maintain, breach of contract). Protects against vendor lock-in and ensures business continuity for critical applications.
💡 Escrow triggers: Vendor insolvency, failure to provide updates/patches, breach of SLA, cessation of business. NOT a substitute for SBOM or code review — escrow covers availability risk, not security quality.

8.4 Software Supply Chain & Code Security

ConceptDefinition
SBOMSoftware Bill of Materials — lists all components/libraries/dependencies
Code SigningDigital signature on code to prove integrity & authenticity
CI/CD Pipeline SecurityAutomated build-test-deploy; integrate SAST, DAST, SCA at each stage
Third-Party LibrariesSCA to identify vulnerable components; pin versions

Maturity Models

ModelLevelsFocus
CMM/CMMI1-Initial → 2-Managed → 3-Defined → 4-Quantitatively Managed → 5-OptimisingProcess maturity
SAMM (OWASP)Governance, Design, Implementation, Verification, OperationsSoftware security maturity
BSIMMObservation-based (descriptive, not prescriptive)Real-world software security practices

8.5 AI/ML Security Considerations

ThreatDescription
Data PoisoningAdversary corrupts training data → model produces wrong results
Adversarial InputsCarefully crafted inputs that fool model (e.g., slightly modified image misclassified)
Model StealingQuerying model enough to reconstruct it
Prompt InjectionManipulating LLM inputs to bypass controls/extract data

Domain 8 — Key Reference Values

OWASP Top 10 (2021) — Priority Order:
A01 Broken Access Control → A02 Cryptographic Failures → A03 Injection → A04 Insecure Design
A05 Security Misconfiguration → A06 Vulnerable & Outdated Components → A07 Authentication Failures
A08 Software & Data Integrity Failures → A09 Security Logging & Monitoring Failures → A10 SSRF

CMMI Maturity Levels:
1-Initial (ad hoc, chaotic) → 2-Managed (basic PM processes) → 3-Defined (standard org process)
4-Quantitatively Managed (metrics-driven control) → 5-Optimising (continuous process improvement)

Database Normalization Quick Test:
1NF: atomic cell values; no repeating groups
2NF: 1NF + no partial key dependencies (every non-key attr depends on entire primary key)
3NF: 2NF + no transitive dependencies (non-key attr does not depend on another non-key attr)

DevSecOps Security Gate Summary:
Commit: SAST → secrets scan → licence check
Build: SCA (dependency vulns) → IaC security scan
Test: DAST → integration security tests → fuzz testing
Deploy: CSPM checks → runtime protection activation → SBOM generation

Domain 8 — Mnemonics

#MnemonicHelps Remember
1"Waterfall = Water flows down only"Can't go back in Waterfall
2"Shift Left = Security Sooner"Test security early in SDLC
3"ACID = All Consistent Isolated Durable"Database transaction properties
4"CMMI 1-5: I Must Define, Quantify, Optimise"CMM maturity levels
5"Parameterised = Protected (from SQLi)"SQL injection prevention
6"The Key, the Whole Key, Nothing but the Key"Normalization: 1NF, 2NF, 3NF
7"Loose Coupling + High Cohesion = Secure Code"OOP design quality
8"Encapsulation = Envelope (hide internals)"Best OOP security feature

Domain 8 — Exam Traps 🚨

#Trap
1SQL injection prevention = parameterised queries, NOT input filtering alone
2XSS prevention = output encoding (not just input validation)
3Buffer overflow prevention = ASLR + DEP + bounds checking
4Aggregation ≠ Inference (aggregation = combining; inference = deducing)
5TOCTOU = race condition; atomic operations fix it
6Waterfall = no iteration; Spiral = risk-driven iteration; Agile = customer-driven iteration
7DevSecOps = "shift left"; integrate security into CI/CD pipeline
8SBOM = know what's in your software (supply chain transparency)
9BSIMM = descriptive (what orgs DO); SAMM = prescriptive (what orgs SHOULD do)
10Fail secure = deny on error; Fail open = allow on error (fail secure is default for security)
11Encapsulation = best OOP security feature (data hiding); inheritance can WEAKEN security (inherits parent's flaws)
12Normalization: 1NF=atomic values, 2NF=no partial dependency, 3NF=no transitive dependency. Denormalization trades integrity for performance.
13Software escrow protects against VENDOR FAILURE, not code quality. Triggers: bankruptcy, cease of support, breach of SLA
14Polyinstantiation prevents inference attacks by creating multiple versions of data at different classification levels
15SAFe = Agile for large enterprises; adds governance layer; exam may distinguish SAFe from Scrum (team-level) and Kanban (flow-based)
16All XSS types are equally dangerous
17CSRF and XSS are the same attack
18Deserialization is safe if input is validated
19Coupling and cohesion are both "higher is better"
20CMMI levels are not test-relevant in CISSP
21Developer testing = acceptance testing
22Authorization checks are only needed at login
23SSRF is a client-side attack

🚨 Master Exam Traps — Consolidated Reference

233 Traps Across All 8 Domains · Night-Before Review · Print Separately from Domain Sections
⚠️ How to use this section: This is a consolidated rapid-fire review — not a replacement for domain-level study. Each trap here also appears in its domain section with full context, reasoning, and memory devices. If a trap makes no sense in isolation, navigate to the relevant domain section for the full explanation. Read this section the night before your exam, after studying each domain. These are the most common deliberate misdirections in CISSP questions — each is a potential 1–2 point swing.

Domain 1 — Security & Risk Management (36 Traps)

#Trap (What the Exam Wants You to Get Wrong)Correct Answer
1Data Owner is IT / sysadminData Owner = business executive; Custodian = IT
2ISO 27002 is certifiableOnly ISO 27001 is certifiable; 27002 is guidance
3"Ignore" as risk responseNot valid. Use "Accept" (documented decision)
4Transfer eliminates accountabilityTransfers financial burden only, NOT legal liability
5Compliance = securityCompliance is the floor, not the ceiling
6Copyright protects ideasCopyright protects expression not ideas
7Trade secrets auto-protectedRequire active protection or they are lost
8GPL is safe for commercialGPL is viral/copyleft — derivative works must be GPL
9EF is a percentage number (50)EF is a decimal (0.50)
10Ransomware primarily hits confidentialityPrimary target = Availability (to extort payment)
11Hot site is always the right answerChoose cheapest option meeting MTD/RTO/RPO
12First action in disaster = restore systemsFirst action = personnel safety
13First action on termination = exit interviewFirst action = revoke access immediately
14SoD prevents collusionSoD prevents fraud, NOT collusion
15RTO = total recovery timeRTO + WRT ≤ MTD; WRT is verification time after restore
16Mandatory vacation is preventiveMandatory vacation = detective (exposes fraud)
17Training completion = effectivenessBehavioural metrics (click/report rate) > activity metrics
18Canon 3 (employer) over Canon 1 (society)Society always first — whistleblowing justified
19Reciprocal agreements are reliableUnreliable — both orgs may face same disaster
20Least Privilege = Need to KnowLP = permissions (what you do); NtK = information (what you see)

Domain 1 — Additional Traps

#❌ Wrong Belief (Trap)✅ Correct Answer
21Higher-numbered canon takes precedence when canons conflictCanon 1 (society/public interest) always overrides higher-numbered canons — whistleblowing is justified when public safety is at stake
22Policy and guideline are both mandatory documentsPolicy = mandatory WHY; Standard = mandatory WHAT; Procedure = mandatory HOW; Guideline = optional HOW — only one non-mandatory level
23Data sovereignty is determined by the cloud provider's HQ countryDetermined by the country where data is stored/processed; provider HQ is irrelevant
24Technical controls are the best fix for human errorRoot cause of human error = lack of awareness; exam answer = security awareness training
25First priority in a disaster is restoring critical systemsPersonnel safety ALWAYS comes first, before any system recovery action
26Work-for-hire copyright = creator's life + 70 yearsIndividual = life + 70 yrs; work-for-hire = 95 yrs from publication OR 120 yrs from creation (whichever is shorter)
27EF is expressed as a percentage (e.g., 50)EF is a decimal (0.50, not 50); ARO is a frequency — 0.1 = once every 10 years
28FAIR is a qualitative risk method like OCTAVEFAIR is quantitative — uses Loss Event Frequency × Loss Magnitude to produce numeric risk values
29Due care and due diligence mean the same thingDifferent concepts: due diligence = research/investigation before acting; due care = ongoing responsible action after deciding
30The CISO is ultimately accountable for organisational securitySenior management (executives/board) are ultimately accountable; CISO is responsible for execution
31Risk appetite and risk tolerance are synonymsDifferent levels: risk appetite = strategic high-level willingness; risk tolerance = tactical acceptable variance from targets
32Delphi technique is structured group brainstormingDelphi = anonymous expert consensus gathered in iterative rounds — specifically designed to prevent groupthink
33BIA and Risk Assessment perform the same analysisDifferent outputs: BIA identifies business impact and criticality (MTD/RTO/RPO); Risk Assessment evaluates threats and likelihood
34Criminal, civil, and administrative law are interchangeable legal categoriesThree distinct categories: Criminal = beyond reasonable doubt, prison/fines; Civil = preponderance of evidence, monetary damages; Administrative = regulatory penalties, licence revocation
35MTD (Maximum Tolerable Downtime) is set by the IT departmentMTD is a business decision made by data owners/business executives, not IT — IT must satisfy the MTD IT cannot define it
36Prudent Person Rule does not apply to cybersecurity professionalsIt applies fully — security professionals are held to the standard of a reasonably prudent person in the same role; negligence is judged against this standard

Domain 2 — Asset Security (22 Traps)

#TrapCorrect Answer
1"Confidential" means the same in govt & commercialLOWEST govt classified but HIGHEST commercial level — naming collision
2Data Owner = IT departmentData Owner = senior management (NOT IT); Custodian = IT
3Degaussing works on SSDsDegaussing SSDs = ZERO effect (data fully intact)
4DoD 5220.22-M overwrite works on SSDsUNRELIABLE for SSDs (wear-levelling bypasses it)
5Use overwrite for cloud data destructionCrypto-erase = preferred for SSDs AND cloud
6Degaussed HDD can be reusedDegaussed HDD becomes permanently inoperable (servo tracks destroyed)
7Retention schedule overrides everythingLitigation hold OVERRIDES all retention schedules
8Anonymised & pseudonymised treated the same under GDPRAnonymised = GDPR no longer applies; Pseudonymised = GDPR still applies
9Processor can do whatever it wants with dataProcessor exceeding instructions BECOMES a controller (full liability)
10Right to Erasure is absoluteNOT absolute — 6 major exceptions (legal claims, public health, etc.)
11DPO can also serve as CISODPO cannot hold CISO/CTO/CEO roles simultaneously (conflict of interest)
12GDPR 72-hour clock starts at incident timeStarts at controller awareness, not incident discovery
13Network DLP sees all trafficNetwork DLP requires SSL inspection to see HTTPS content
14Data in transit is hardest to protectData in use is HARDEST to protect (requires hardware TEE)
15FDE protects a running machineFDE useless on powered-on, logged-in devices
16Mixed-classification system = lowest levelClassify UP: system classified at the highest level of data it handles
17Simple delete = secure destructionStandard file deletion only removes pointers — data remanence persists. Use NIST 800-88.
18EOS systems are fine if isolatedEOS = unpatched = high-risk. Must: isolate, increase monitoring, compensating controls, document risk acceptance.
19Data steward = data ownerSteward handles data quality/governance; owner sets classification/access decisions
20Scoping = tailoringScoping selects controls; tailoring customizes selected controls
21Marking and labeling are the sameMarking is internal/metadata context; labeling is external/physical indicator
22All sanitization methods are equivalentNIST 800-88: Clear < Purge < Destroy; choose by sensitivity and reuse intent

Domain 3 — Security Architecture & Engineering (35 Traps)

#TrapCorrect Answer
1BLP and Biba have the same rulesEXACT opposites. BLP = confidentiality (no read up/write down); Biba = integrity (no read down/write up)
2Clark-Wilson is about confidentialityClark-Wilson = commercial integrity (well-formed transactions via TPs on CDIs)
3Brewer-Nash uses static rulesDynamic rules based on access HISTORY (Chinese Wall)
4SABSA / Zachman / TOGAF are all the sameSABSA = security framework; Zachman/TOGAF = enterprise architecture (not security-specific)
5EAL 7 is the target for commercial productsEAL 4 = highest practically achievable; EAL 7 = formal verification (extremely rare)
6Reference Monitor = Security KernelReference Monitor = concept; Security Kernel = implementation
7Data leakage between VMs is the top threatVM Escape is the most critical virtualisation attack
8SCADA/ICS uses CIA priority orderSCADA priority = Safety first, NOT confidentiality (SAIC not CIA)
9AES block size changes with key sizeAES block size ALWAYS 128 bits — only key size varies (128/192/256)
10ECB mode is acceptableECB is INSECURE — identical plaintext blocks → identical ciphertext blocks (patterns leak)
11DH and DSA can do everything RSA doesDH = key exchange ONLY; DSA = signatures ONLY; RSA = everything
12HMAC provides non-repudiationHMAC cannot provide non-repudiation (shared secret key — both parties could generate)
13Encryption provides non-repudiationDigital signatures = ONLY way to get non-repudiation
14128-bit hash = 128-bit collision resistanceBirthday attack = 2^(n/2); 128-bit hash = only 64-bit collision resistance
15RA can sign certificatesRA verifies identity but does NOT sign certificates
16Root CA should be online for availabilityRoot CA must be OFFLINE (air-gapped)
17Escrow all keys for recoveryNEVER escrow signing keys (destroys non-repudiation). Only escrow encryption keys.
18Side-channel attacks break the algorithmSide-channel attacks exploit implementation, NOT algorithm math
19TLS 1.3 supports RSA key exchangeTLS 1.3 mandates PFS (only DHE/ECDHE); RSA key exchange removed
20Containers have the same isolation as VMsContainer isolation WEAKER than VM (shared kernel — kernel compromise = all containers)
21Dedicated security mode is the most complexDedicated = simplest; Multilevel = most complex (different clearances; needs MAC)
22Covert storage and timing channels are equally hard to detectCovert timing channels are HARDER to detect than storage channels
23RAM is empty immediately after power-offCold boot attack: RAM retains data briefly — freeze chips to extract encryption keys
24Private cloud = on-premisesPrivate cloud can be hosted off-site by a third party — key is exclusive use
25TCSEC B2 requires covert channel mitigationB2+ requires covert channel ANALYSIS; B3+ requires MITIGATION
26Encapsulation is just about hiding codeEncapsulation = BEST OOP security feature (data hiding). Inheritance can weaken security.
27ECC needs the same key size as RSAECC-256 ≈ RSA-3072 ≈ 128-bit security level. ECC punches above its weight.
28Any block cipher mode provides authenticationOnly GCM (AEAD) provides both confidentiality AND authentication. CBC/CTR do NOT.
29Symmetric key count = 2nSymmetric = n(n−1)/2; Asymmetric = 2n. Don't mix them up.
30Security depends on keeping algorithm secretKerckhoffs: security should depend on key secrecy, not algorithm secrecy
31Quantum breaks all encryption equallyAsymmetric is hit hardest; AES-256 remains viable with adjusted assumptions
32One-time pad is always unbreakable in real deploymentsOnly if key is truly random, never reused, and at least message-length
33Zero-knowledge proof requires revealing secretIt proves knowledge without disclosing the secret
34Stream and block ciphers are interchangeableDifferent primitives; do not confuse cipher type with mode of operation
35Confusion and diffusion are the sameConfusion obscures key relation; diffusion spreads plaintext influence

Domain 4 — Communication & Network Security (23 Traps)

#TrapCorrect Answer
1IPv6 requires IPSec useIPv6 has mandatory IPSec SUPPORT, not mandatory USE
2DNSSEC encrypts DNS queriesDNSSEC = integrity/authentication only, NOT confidentiality (use DoH/DoT for that)
3AH provides encryptionAH provides NO encryption; also broken by NAT (NAT modifies IP header)
4TLS 1.3 is backward-compatible with RSA key exchangeTLS 1.3 = 1-RTT, mandates PFS, removed RSA key exchange
5WEP is weak but usableWEP IV is only 24 bits → exhausted in minutes → trivially cracked
6WPA3 and WPA2 handle offline attacks the sameWPA3-SAE resists offline dictionary attacks (unlike WPA2-PSK)
7802.1X uses TACACS+802.1X uses RADIUS (not TACACS+) for wireless enterprise auth
8Switches are inherently secureSwitches can be compromised (MAC flooding, VLAN hopping)
9IPSec VPN is always easier to deploy than SSL VPNSSL VPN works through firewalls (port 443) — advantage over IPSec
10SDN has distributed controlSDN separates control plane from data plane (centralised management)
11Zero Trust = block everythingZero Trust = "never trust, always verify" not "block everything"
12Smurf = TCP attackSmurf attack uses ICMP + broadcast + spoofed source (amplification)
13Management plane doesn't need special protectionManagement plane = most sensitive; must isolate via out-of-band management
14Cut-through switching is more secureCut-through is faster but passes corrupted frames; store-and-forward checks CRC (more secure)
15Data sovereignty = country of company HQData sovereignty = governed by country where data is STORED
16Multilayer protocols are always inspected by security controlsEncapsulated protocols may bypass security controls that only inspect one layer (VPN within VPN, IPv6 in IPv4)
17Bluetooth attacks are all equivalentBluejacking (message), bluesnarfing (data theft), bluebugging (device control)
18FTPS and SFTP are the same protocolFTPS = FTP over TLS; SFTP = SSH file transfer protocol
19L2TP encrypts traffic by itselfL2TP needs IPSec for encryption
20SYN floods are difficult to mitigateSYN cookies and rate limiting are standard controls
21S/MIME and PGP use the same trust modelS/MIME is CA/PKI; PGP is web-of-trust
22Fraggle and Smurf are identicalSmurf uses ICMP; Fraggle uses UDP echo amplification
23DNS cache poisoning affects one host onlyCompromised resolver can poison responses for all its clients

Domain 5 — Identity & Access Management (25 Traps)

#TrapCorrect Answer
1Two passwords = MFANOT MFA (same factor type); must be DIFFERENT factor types
2Type 1 error = authentication Type 1 factorType 1 error = False Rejection (FRR); Type 2 error = False Acceptance (FAR). Don't confuse with auth factor types.
3OAuth 2.0 authenticates usersOAuth 2.0 is AUTHORISATION only — use OIDC for authentication
4SAML and OIDC use the same formatSAML uses XML; OIDC uses JWT tokens (modern, mobile-friendly)
5Kerberos uses asymmetric cryptoKerberos = symmetric crypto + time-dependent; Golden Ticket = compromised KRBTGT
6RADIUS encrypts everythingRADIUS encrypts password only; TACACS+ encrypts entire payload
7RBAC and Rule-Based are the sameRBAC ≠ Rule-Based. RBAC = Role-Based; Rule-Based = IF-THEN (e.g., firewall rules)
8Iris scan is the most invasiveRetina scan = most accurate AND most invasive; can reveal health info
9Access rights only grow when appropriatePrivilege creep = accumulation without revocation. Fix with periodic UARs.
10MAC and DAC have similar restrictivenessDAC is most permissive (owner decides); MAC is most restrictive (system decides)
11Federation and SSO are the sameFederation = cross-org trust; SSO = same organisation, multiple apps
12Smart card is single-factorSmart card = possession (Type 2) + PIN = knowledge (Type 1) = MFA
13RBAC and risk-based access control work the same wayRisk-based = DYNAMIC (changes per session); RBAC = STATIC (changes per role assignment)
14Session tokens can go in URLsSession tokens must be regenerated after login (prevent fixation); never in URLs (bookmarks/logs leak them)
15Service accounts are low-riskService accounts = top attack vector; hardcoded creds, orphan accounts, no rotation = audit failures
16Service accounts need stored passwordsManaged identities (cloud-native) eliminate stored credentials entirely — preferred approach
17FAR is less dangerous than FRRFAR is MORE DANGEROUS (lets intruder in). CER/EER = where FRR=FAR (lower = more accurate).
18Kerberos has unlimited clock toleranceKerberos requires 5-minute clock skew tolerance (NTP synchronisation critical)
19Push-based MFA is phishing-resistantPush-based MFA alone is NOT phishing-resistant. MFA fatigue/push bombing is real. Use FIDO2/WebAuthn or number matching.
20Identity proofing = authenticationProofing is initial enrollment identity verification; authentication is ongoing login verification
21RBAC is the most granular modelABAC is generally more granular and context-aware
22Privileged users should have standing accessJIT/JEA reduces standing privilege risk
23PAM and IAM are the same scopePAM is a focused subset for privileged identities
24Password spraying = brute force = credential stuffingDistinct attack patterns with different detection/mitigation
25Authentication only happens at loginContinuous/adaptive auth validates session posture over time

Domain 6 — Security Assessment & Testing (18 Traps)

#TrapCorrect Answer
1Vulnerability scan = penetration testVulnerability scan identifies; penetration test exploits
2Verbal approval is sufficient for pentestWritten authorisation REQUIRED before any pentest (no verbal OK)
3SOC 2 Type I and Type II are equally valuableSOC 2 Type II > Type I for ongoing assurance (period vs point-in-time)
4Code review catches all bugsFuzz testing finds bugs code review misses (crashes, memory corruption)
5MTBF and MTTF are interchangeableMTBF includes repair time (repairable items); MTTF = until first failure (no repair assumed)
6Logs are tamper-proof by defaultLog integrity requires centralised logging + tamper protection + NTP
7Red team and Purple team are the sameRed team = stealth adversary sim; Purple team = collaborative improvement
8SCA scans your codeSCA finds vulnerable open-source libraries (not your code, their bugs)
9KPI and KRI are interchangeableKPI = performance (backward-looking); KRI = risk warning (forward-looking)
10Backup logs confirming success = backups workBackup verification = actually TEST restores regularly, not just check logs
11SOC 2 can skip Security criteriaSecurity (Common Criteria) is MANDATORY in every SOC 2 report. Other TSC categories are optional.
12SAST and DAST test the same thingSAST = static source code analysis (white-box); DAST = running app testing (black-box). Both needed.
13Synthetic monitoring = Real User MonitoringSynthetic = scripted/proactive (pre-production); RUM = actual users/passive (production)
14SOC 1, SOC 2, SOC 3 are equivalent reportsSOC 1 = financial controls; SOC 2 = trust services; SOC 3 = public summary
15100% code coverage proves software is secureCoverage is a quantity metric, not proof of defect absence
16Patches do not require regression testingRegression testing is required to avoid break/fix side effects
17False positives are more dangerous than false negativesFalse negatives are riskier because true attacks are missed
18Only use-case testing is neededAbuse/misuse-case testing is critical for security assurance

Domain 7 — Security Operations (38 Traps)

#TrapCorrect Answer
1First priority in incident = contain the breachFIRST = human safety, THEN containment, THEN evidence
2Evidence is always admissibleEvidence without chain of custody = INADMISSIBLE
3Image hard drive firstCollect most volatile evidence FIRST (RAM before disk)
4Hash forensic image after creation onlyForensic images must be hashed before AND after (integrity proof)
5RAID = backupRAID is NOT backup — protects against disk failure only, not data loss/corruption
6RAID 0 has some redundancyRAID 0 = performance only; 1 disk fails = ALL data lost
7Incremental backup = fast restoreIncremental restore needs Full + ALL incrementals (slow restore)
8Differential backup = slow restoreDifferential restore needs Full + LAST differential only (faster restore)
9CO₂ is safe for data centresCO₂ fire suppression = lethal to humans; FM-200 = safe for people
10Halon is still availableHalon is banned (ozone depletion); FM-200 is the replacement
11Criminal and civil cases use the same burden of proofCriminal = beyond reasonable doubt; Civil = preponderance of evidence
12Emergency changes skip CABEmergency changes have expedited approval but still require retrospective CAB review
13CCTV prevents tailgatingCCTV detects but doesn't prevent. Mantrap/vestibule prevents tailgating.
14Lessons learned can wait monthsLessons learned meeting ideally within 1-2 weeks of incident
15Virus and worm are the sameVirus needs host + user action; Worm self-propagates across network autonomously
16OS reinstall fixes all rootkitsOS rootkit = full reinstall; firmware rootkit = REPLACE HARDWARE
17Polymorphic and metamorphic are the samePolymorphic changes signature; metamorphic rewrites entire code — both evade signature AV
18Blocklisting is effective enoughAllowlisting > blocklisting (default-deny is stronger than default-allow)
19Honeypot = entrapmentEnticement = legal (honeypot lures existing intent); Entrapment = illegal (inducing crime)
20Water works for electrical firesFire Class C = electrical — NEVER use water (electrocution risk); use CO₂ or FM-200
21DR plan only needs one communication channelDR Communications plan must have BACKUP channels (primary comms may be destroyed)
22Cloud forensics is like traditional forensicsCloud: SaaS = least forensic access; IaaS = most. Negotiate rights in contract BEFORE incident.
23Technical controls are best for insider threatsUEBA = best technical control (baselines behaviour, detects anomalies). But combine with administrative controls (rotation, vacation).
24Manual config management is fineIaC + config automation (Ansible/Puppet) enforces baseline at scale — manual doesn't work for large environments
25Recovery = restorationRecovery = move TO DR site; Restoration = move BACK to primary. During restoration, DR site becomes the backup.
26Push-based MFA is fully secureMFA fatigue/push bombing: attacker spams push notifications. Fix: number matching, FIDO2, rate limiting.
27Low humidity is fine for electronicsToo low (<40%) → ESD damage; too high (>60%) → condensation/corrosion. Ideal: 40-60% RH.
28Spoliation has minor consequencesFailure to preserve evidence under litigation hold = spoliation = severe legal sanctions
29Egress monitoring is optionalEgress filtering is critical but often neglected. DLP + firewall + proxy prevent data exfiltration and C2 callbacks.
30UPS is sufficient for extended outagesUPS = short-term (minutes); Generator = long-term. Blackout needs BOTH.
31Evidence copies are always equal to originals in courtOriginal preferred; copies need authenticity proof (hash + custody)
32System logs are always inadmissible hearsayCan be admitted under business-records exception with proper controls
33All evidence types carry equal legal weightReal/documentary/testimonial/demonstrative have different strengths
34Criminal burden of proof applies to all casesStandards vary by criminal, civil, and administrative context
35IDS alone detects all attacksSignature misses unknowns; anomaly catches unknowns with more false positives
36SPAN ports are as forensically reliable as tapsTaps are more reliable; SPAN can drop packets under load
37Clipping levels are optional tuning noiseThreshold tuning is essential to avoid alert fatigue or misses
38Warm site = hot site without dataRecovery capability and activation effort differ significantly by hot/warm/cold

Domain 8 — Software Development Security (26 Traps)

#TrapCorrect Answer
1Input filtering stops SQL injectionSQL injection prevention = parameterised queries, NOT input filtering alone
2Input validation stops XSSXSS prevention = output encoding (not just input validation)
3Buffer overflows need one defenceBuffer overflow prevention = ASLR + DEP + bounds checking (defence in depth)
4Aggregation and inference are the sameAggregation = combining; Inference = deducing
5Race conditions are hard to fixTOCTOU = race condition; fix with atomic operations and locks
6Waterfall/Spiral/Agile are interchangeableWaterfall = no iteration; Spiral = risk-driven iteration; Agile = customer-driven iteration
7Security can be added at the end of the SDLCDevSecOps = "shift left"; integrate security into CI/CD pipeline from the start
8Supply chain is someone else's problemSBOM = know what's in your software (supply chain transparency)
9BSIMM and SAMM are the sameBSIMM = descriptive (what orgs DO); SAMM = prescriptive (what orgs SHOULD do)
10Fail open is the secure defaultFail secure = deny on error; Fail open = allow on error. Fail secure is the security default.
11All OOP features improve security equallyEncapsulation = best OOP security feature (data hiding); inheritance can WEAKEN security (inherits parent's flaws)
12Denormalization is always badNormalization: 1NF=atomic, 2NF=no partial dep, 3NF=no transitive dep. Denormalization trades integrity for performance (intentional choice).
13Software escrow = code reviewSoftware escrow protects against VENDOR FAILURE, not code quality. Triggers: bankruptcy, SLA breach.
14Inference attacks can't be preventedPolyinstantiation prevents inference attacks by creating multiple data versions at different classification levels
15SAFe = ScrumSAFe = Agile for large enterprises; adds governance layer. Scrum = team-level. Kanban = flow-based.
16API keys = authenticationAPI keys = identification only, NOT authentication. Use OAuth 2.0/OIDC for API auth.
17OWASP Top 10 Web and API are the sameOWASP API Security Top 10 is separate. Key API risks: BOLA (Broken Object Level Auth), mass assignment, excessive data exposure.
18Whitelist (allow-list) input validation < blacklistWhitelist (allow known-good) > Blacklist (block known-bad) — always prefer whitelist approach
19All XSS variants are equally dangerousStored XSS usually has broader impact than reflected/DOM depending on context
20CSRF and XSS are the same attack classCSRF forges trusted requests; XSS injects script into trusted content
21Deserialization is safe if basic validation existsUnsafe deserialization can trigger RCE; avoid untrusted object deserialization
22Coupling and cohesion both should be highPrefer low coupling and high cohesion
23CMMI levels are not exam-relevantKnow progression from Initial to Optimizing
24Developer test completion = user acceptanceAcceptance testing is business/user validation against requirements
25Authorization checks are only needed at loginEnforce object-level authorization on every request (IDOR prevention)
26SSRF is a client-side issueSSRF coerces server-side requests to internal/unintended targets

Cross-Domain Mega-Traps (Tested Across Multiple Domains)

#TrapDomains
1"Think like a manager, not a technician." CISSP is a management exam. When two answers are technically correct, choose the one that addresses governance, risk, policy, or process over the technical fix.ALL
2Human safety is ALWAYS the first priority. Before recovering systems, saving evidence, or containing the threat — protect people.D1, D7
3Data Owner = business executive. ALWAYS. Not IT, not DBA, not CISO. The owner is the person who classifies the data and approves access. Appears in D1, D2, D5.D1, D2, D5
4Compliance ≠ Security. You can be compliant and still get breached. Compliance is the floor. Security goes beyond.D1, D6
5Non-repudiation requires digital signatures. Not encryption, not hashing, not HMAC. Only asymmetric digital signatures provide non-repudiation. Appears in D3, D5.D3, D5
6Least privilege ≠ Need-to-know. LP = what you CAN DO (permissions). NtK = what you CAN SEE (information). Both needed together.D1, D5, D7
7"FIRST" / "BEST" / "MOST IMPORTANT" — Read these words very carefully. They narrow the answer to exactly ONE choice. Eliminate good-but-not-best answers.ALL
8Legal liability cannot be transferred. You can transfer financial risk (insurance) but you REMAIN legally accountable. Risk transfer ≠ accountability transfer.D1, D2
9The answer CISSP wants is the PROCESS answer, not the product. "Implement a risk assessment" beats "buy a firewall." "Develop a policy" beats "install antivirus."ALL
10When in doubt, pick the answer that reduces risk to the organisation. Not the cheapest, not the most technical, not the fastest — the one that best manages risk within business context.ALL
📊 TOTAL EXAM TRAPS: 233 (D1: 36 · D2: 22 · D3: 35 · D4: 23 · D5: 25 · D6: 18 · D7: 38 · D8: 26 · Cross-Domain: 10). Review this section the night before your exam. If you get every trap right, you're well above the pass line.

Context & Cross-Domain Study Note

This section lists what to know — the domain sections explain why. Context makes answers stick. Every trap listed here is explained in full within its domain’s “Exam Traps” subsection and surrounding concept sections — that is where you’ll find the reasoning, examples, and memory devices that make the correct answer memorable.

If a trap is unclear in isolation: Find the relevant domain section using the links below and review the concept that surrounds it. The “See also” boxes throughout the domain sections highlight cross-domain connections that are frequently combined in exam questions.

Exam mindset summary: CISSP questions reward thinking like a security manager, not a technician. When two answers look equally valid, ask: which one better serves the organisation’s risk management objectives? Which aligns with a documented, approved process? Which would a prudent senior CISSP professional choose — and be comfortable defending to their board?

Quick-navigate to domain sections:
D1: Risk & Governance (16%) · D2: Asset Security (10%) · D3: Architecture (13%) · D4: Network Security (13%) · D5: IAM (13%) · D6: Assessment (12%) · D7: Operations (13%) · D8: Software Security (10%)

📐 Quick Formulas & Numeric Reference — All Domains

Consolidated Print Reference · All Key Numbers in One Place
Each formula also appears in its domain section with full context and explanation. This appendix is a rapid-reference summary only — use it for last-minute review once you understand the underlying concepts.

D1 — Risk Management

SLE = Asset Value × EF   (EF = decimal 0.0–1.0)
ALE = SLE × ARO
Control value = ALEbefore − ALEafter − annual control cost

Recovery timelines:
RTO + WRT ≤ MTD
• RPO = data loss tolerance (backup frequency)
• RTO = system must be operational by
• WRT = post-recovery validation time
• MTD = max tolerable total downtime

D3 — Cryptography

Symmetric keys needed: n(n−1)/2
Asymmetric key pairs: 2n
Birthday attack: 2n/2 (n-bit hash)
Brute-force space: 2n (n-bit key)

AES block: always 128 bits
DES: 56-bit key · 64-bit block · 16 rounds
ECC-256 ≈ RSA-3072 ≈ 128-bit security
ECC-384 ≈ RSA-7680 ≈ 192-bit security

D5 — Identity & Access

FRR = Type 1 Error (legitimate user rejected)
FAR = Type 2 Error (impostor accepted — more dangerous)
CER/EER = crossover point (FRR = FAR; lower = better)
Decreasing FAR → increases FRR (trade-off)

Kerberos clock skew: 5 min
Kerberos TGT lifetime: 10 hours

D6 — Assessment & CVSS

CVSS v3.x score ranges:
0.0 None · 0.1–3.9 Low · 4.0–6.9 Medium
7.0–8.9 High · 9.0–10.0 Critical

MTBF = MTTF + MTTR   (repairable systems)
MTTD = time to detect
MTTR = time to remediate
MTTF = time to first failure (non-repairable)

D7 — Operations & Recovery

Availability = MTBF ÷ (MTBF + MTTR) × 100%

Uptime benchmarks:
99% = 3.65 days/yr · 99.9% = 8.77 hrs/yr
99.99% = 52.6 min/yr · 99.999% = 5.26 min/yr

RAID usable capacity:
RAID 0: n×disk · RAID 1: 1×disk
RAID 5: (n−1)×disk · RAID 6: (n−2)×disk
RAID 10: (n/2)×disk

Breach notifications:
GDPR → Authority: 72 hours
HIPAA → HHS: 60 days
PCI-DSS → card brands: immediately

D8 — Software Security

OWASP Top 3 (2021):
A01 Broken Access Control
A02 Cryptographic Failures
A03 Injection

CMMI Levels:
1 Initial → 2 Managed → 3 Defined
4 Quantitative → 5 Optimising

Normalisation:
1NF: atomic values
2NF: no partial key deps
3NF: no transitive deps