πŸ›οΈ CISSP Domain 3 Β· Security Architecture & Engineering

Bell-LaPadula Β· Biba Β· Clark-Wilson Β· Brewer-Nash
Security Models β€” 2026 CISSP Guide

The most visual, interactive guide to CISSP security models. Understand the rules, spot the signals, and never confuse these models on exam day.

πŸ”΄ Confidentiality
πŸ”΅ Integrity (lattice)
🟒 Integrity (transactions)
🟣 Conflict of Interest
OVERVIEW

What Are Security Models?

Security models are formal frameworks that define rules for how subjects (users/processes) interact with objects (data/resources). They underpin OS and system design and are heavily tested on the CISSP exam as scenario-based questions.

πŸ”΄

Bell-LaPadula

Lattice-Based Β· 1973

Protects confidentiality. Information can only flow upward or stay level β€” never down. Built for US military classification systems.

πŸ”’ Protects: CONFIDENTIALITY
πŸ”‘ "No Read Up Β· No Write Down"
πŸ”΅

Biba

Lattice-Based Β· 1977

Protects integrity. Low-integrity data cannot contaminate high-integrity data. The exact inverse of Bell-LaPadula.

βœ… Protects: INTEGRITY
πŸ”‘ "No Write Up Β· No Read Down"
🟒

Clark-Wilson

Transaction-Based Β· 1987

Protects integrity through well-formed transactions and separation of duties. Designed for commercial environments.

βœ… Protects: INTEGRITY
πŸ”‘ "Procedures + Separation of Duties"
🟣

Brewer-Nash

Chinese Wall Β· 1989

Prevents conflict of interest. Access is dynamically restricted based on what data the user has already accessed.

🧱 Prevents: CONFLICT OF INTEREST
πŸ”‘ "Dynamic Wall Β· Prior Access History"
πŸ’‘
CISSP Exam Tip: Ask yourself β€” what does this scenario protect? Confidentiality + labels β†’ Bell-LaPadula. Integrity + levels β†’ Biba. Integrity + approved procedures + separation of duties β†’ Clark-Wilson. Conflict of interest + dynamic restriction β†’ Brewer-Nash.

Quick Reference Snapshot

Model Protects Read Rule Write Rule Access Type
Bell-LaPadulaConfidentiality❌ No Read Up❌ No Write DownStatic (labels)
BibaIntegrity❌ No Read Down❌ No Write UpStatic (levels)
Clark-WilsonIntegrityVia IVP onlyVia TP + SoD onlyProcedural
Brewer-NashConf. + IntegrityBased on historyBased on historyDynamic (wall)
⚠️
Critical exam trap β€” BLP vs Biba: Bell-LaPadula says "No Write Down" (can't leak secrets to lower levels). Biba says "No Write Up" (can't corrupt higher-integrity data). They are exact opposites. If you mix these up on the exam, you'll lose easy points. Use the How It Works tab to see the visual difference.
HOW IT WORKS

Rules & Access Flow Diagrams

Select a model to see exactly how its rules work β€” with visual level towers and flow diagrams.

πŸ”΄ Bell-LaPadula Model

Developed by the US DoD in 1973. Enforces confidentiality by ensuring classified information can only flow upward β€” never down to lower-clearance systems.

πŸ“–
Simple Security Property
❌ No Read Up (NRU)
A subject cannot read objects at a HIGHER classification level than their clearance.
✏️
Star Property (*-property)
❌ No Write Down (NWD)
A subject cannot write to objects at a LOWER classification level. Prevents leaking secrets downward.

Access Tower β€” Subject at SECRET Level

⬆️ TOP SECRET
❌ Cannot Readβœ… Can Write
πŸ‘€ SECRET ← Subject Here
βœ… Can Readβœ… Can Write
⬇️ CONFIDENTIAL
βœ… Can Read❌ Cannot Write
⬇️ UNCLASSIFIED
βœ… Can Read❌ Cannot Write
🎯
Why "Can Write Up"? Writing UP (to a higher classification) is allowed β€” it's considered safe. Adding SECRET information into a TOP SECRET document doesn't leak anything. But writing DOWN (to UNCLASSIFIED) would be leaking classified info β€” that's exactly what BLP prevents.

Key Bell-LaPadula Characteristics

  • πŸ”’
    Protects: Confidentiality only β€” Bell-LaPadula does NOT address data integrity
  • πŸ›οΈ
    Origin: Created for the US Department of Defense to secure classified military systems
  • πŸ“‹
    Third rule: Discretionary Security Property β€” uses an access matrix alongside mandatory rules
  • ⚠️
    Weakness: Ignores data integrity β€” a high-clearance user could corrupt data but BLP won't stop it
  • 🎯
    CISSP signals: "classified," "clearance level," "no read up," "no write down," "confidentiality," "military system"

πŸ”΅ Biba Integrity Model

Developed by Kenneth Biba in 1977. Enforces integrity by preventing low-integrity data from contaminating high-integrity systems. The direct inverse of Bell-LaPadula.

πŸ“–
Simple Integrity Property
❌ No Read Down (NRD)
A subject cannot read from objects at a LOWER integrity level β€” contamination risk.
✏️
Star Integrity Property (*)
❌ No Write Up (NWU)
A subject cannot write to objects at a HIGHER integrity level. Prevents corrupting trusted data.

Integrity Tower β€” Subject at MEDIUM Level

⬆️ HIGH INTEGRITY
βœ… Can Read❌ Cannot Write
πŸ‘€ MEDIUM ← Subject Here
βœ… Can Readβœ… Can Write
⬇️ LOW INTEGRITY
❌ Cannot Readβœ… Can Write
🎯
Why "Can Read Up"? Reading HIGH-integrity data is safe β€” it won't contaminate the subject. But reading LOW-integrity data could cause the subject to act on untrustworthy information. Writing UP would corrupt trusted data. Writing DOWN is allowed β€” degrading your own data doesn't compromise the high-integrity system.

Key Biba Characteristics

  • βœ…
    Protects: Integrity only β€” Biba does NOT address confidentiality
  • πŸ”„
    Relationship to BLP: Biba is the mathematical inverse of Bell-LaPadula β€” read/write rules are exactly reversed
  • 🏦
    Real-world use: Financial transaction systems, air traffic control, medical records, any system where data accuracy is paramount
  • ⚠️
    Weakness: Impractical in many real environments β€” users often need to read lower-integrity data (emails, reports) without contaminating their work
  • 🎯
    CISSP signals: "integrity level," "contamination," "unverified source," "no write up," "no read down," "trust level"

🟒 Clark-Wilson Integrity Model

Developed in 1987 for commercial use. Protects integrity through controlled transactions and separation of duties β€” not level-based rules like Biba.

πŸ“¦
CDI
Constrained Data Items
Protected, high-value data that can ONLY be modified through approved TPs (e.g., account balances, medical records)
πŸ“„
UDI
Unconstrained Data Items
Uncontrolled input data (e.g., forms, raw feeds) β€” must be processed before becoming CDI
πŸ”§
TP
Transformation Procedures
The ONLY approved way to modify CDIs. No direct editing allowed β€” all changes go through TPs
βœ…
IVP
Integrity Verification Procedures
Checks that CDIs are valid and consistent. Runs periodically to confirm data integrity

Clark-Wilson Process Flow

πŸ“„
UDI
Raw input data (unverified)
β†’
βœ…
IVP
Integrity verification
β†’
πŸ“¦
CDI
Protected data (constrained)
β†’
πŸ”§
TP
Approved procedure only
β†’
πŸ“¦
CDI Updated
Integrity preserved
🏦
Think: Double-entry bookkeeping. In accounting, every transaction follows an approved procedure (TP), the ledger (CDI) can never be directly edited, an auditor verifies integrity (IVP), and the person who records a transaction cannot also approve it (separation of duties).

Key Clark-Wilson Characteristics

  • βœ…
    Protects: Data integrity through procedural controls and separation of duties
  • 🏒
    Designed for: Commercial environments β€” accounting, ERP systems, healthcare workflows, not military
  • πŸ‘₯
    Separation of duties: No single person can create AND approve a transaction β€” enforced by the model itself
  • πŸ“
    Access triple: (Subject, TP, CDI) β€” a user can only touch a CDI via a specific TP they are authorized to use
  • 🎯
    CISSP signals: "separation of duties," "well-formed transaction," "approved procedure," "CDI," "TP," "no direct modification," "accounting," "dual authorization"

🟣 Brewer-Nash (Chinese Wall) Model

Developed in 1989. Prevents conflict of interest by dynamically restricting access based on what the user has already accessed. The wall grows as access history builds.

🧱
The Chinese Wall Rule
Once you access Company A's data, you are permanently walled off from all competitors of Company A
Access history drives future access decisions. No override possible β€” not even by management. The wall grows as the user accesses more company data.

The Wall in Action

βœ… Accessed: RetailCo (Class A)
πŸ“ RetailCo Financial Data
πŸ“ RetailCo Strategy Docs
πŸ“ RetailCo M&A Files
User has accessed this dataset
🧱
CHINESE WALL BUILT
🚫
❌ Blocked: ShopMart (Class A)
πŸ“ ShopMart Financial Data
πŸ“ ShopMart Strategy Docs
πŸ“ ShopMart M&A Files
Competitor β€” access denied
🏦
Different conflict classes: The wall only applies within a conflict class (competitors). Companies in completely different industries are in different conflict classes and remain accessible. A consulting firm analyst blocked from RetailCo competitors can still access TechCorp or BankCo if those are different conflict classes.

Key Brewer-Nash Characteristics

  • πŸ”„
    Dynamic access: Access rules change over time based on the user's access history β€” unlike all other models which are static
  • 🧱
    No override: Management, directors, or senior partners cannot grant access across a Chinese Wall β€” the restriction is absolute
  • πŸ›οΈ
    Designed for: Consulting firms, law firms, investment banks, financial analysts β€” anywhere conflict of interest is a legal or ethical concern
  • πŸ“š
    Conflict classes: Companies are grouped into "conflict of interest classes." The wall applies within the class, not across classes
  • 🎯
    CISSP signals: "conflict of interest," "Chinese Wall," "consulting firm," "law firm," "previously accessed," "dynamic restriction," "competitor data," "grows over time"
COMPARISON TABLE

Side-by-Side Comparison

Filter by category to focus your study. Hover any row for highlighting.

Filter:
Criteria πŸ”΄ Bell-LaPadula πŸ”΅ Biba 🟒 Clark-Wilson 🟣 Brewer-Nash
Primary goalConfidentialityIntegrityIntegrityConflict of interest prevention
Read rule❌ No Read Up
Can't read above clearance
❌ No Read Down
Can't read lower integrity
Via IVP only
Verification before access
Based on access history
Wall grows dynamically
Write rule❌ No Write Down
Can't leak to lower level
❌ No Write Up
Can't corrupt higher level
Via TP only
Approved procedures only
Based on conflict class
No cross-competitor writes
Access typeStatic (lattice-based)Static (lattice-based)Static (procedural)βœ… Dynamic (history-based)
Addresses CIAConfidentiality onlyIntegrity onlyIntegrity (+ some C via SoD)Both C and I
Designed forMilitary / GovernmentCommercial integrityCommercial / BusinessFinance / Consulting / Legal
Real-world examplesNSA systems, JWICS, military classified networksAir traffic control, financial ledgers, medical systemsAccounting software, ERP (SAP), hospital EMR workflowsLaw firms, consulting (Deloitte, McKinsey), investment banks
Separation of duties❌ Not required❌ Not requiredβœ… Core requirement⚠️ Implicit (via wall)
Can management override?❌ No❌ No⚠️ Limited (must follow TP)❌ No β€” wall is absolute
Often confused withBiba (rules are inverse)Bell-LaPadula (rules are inverse)Biba (both = integrity)RBAC (both restrict access)
Key differentiatorStops secrets leaking DOWNStops contamination flowing UPIntegrity via PROCEDURES not levelsAccess based on HISTORY not levels
CISSP signal words"clearance," "classified," "no read up," "no write down""integrity level," "contamination," "no write up," "no read down""separation of duties," "TP," "CDI," "well-formed transaction""conflict of interest," "Chinese Wall," "previously accessed," "dynamic"
REAL-WORLD EXAMPLES

See Each Model in Action

Click any example to expand the full scenario, why it fits the model, and the CISSP exam takeaway.

Bell-LaPadulaπŸ›οΈ US Intelligence Community Networks
β–Ό
Protects
Confidentiality
Key Rule
No Read Up / No Write Down
Environment
Military / Government

A CIA analyst holds a SECRET clearance. On JWICS (the classified government network), they can read SECRET and CONFIDENTIAL documents freely. When they attempt to read a TOP SECRET intelligence report, the system denies access β€” they haven't been cleared to that level. Later, when drafting their analysis, the system prevents them from saving the document to an UNCLASSIFIED folder, even by accident. The analyst can only save to SECRET or above. This prevents classified information from ever reaching lower-clearance systems, even through human error.

🎯
CISSP Takeaway: Bell-LaPadula's genius is protecting against accidental as well as deliberate leakage. The "No Write Down" rule stops mistakes β€” even a TOP SECRET analyst can't accidentally leak to an UNCLASSIFIED folder. If you see "clearance" + "classification labels" + "downward leakage prevention," it's BLP.
Bell-LaPadulaπŸ’» SELinux Type Enforcement
β–Ό
Protects
Confidentiality
Key Rule
System enforced labels
Environment
Linux OS Security

SELinux implements Bell-LaPadula principles through Multi-Level Security (MLS). Processes are assigned security levels and the kernel enforces read/write restrictions between levels. A web server process (LOW level) cannot read database credential files (HIGH level). Even root-level processes respect these labels. This is exactly BLP's Simple Security Property and *-property implemented in OS kernel code.

🎯
CISSP Takeaway: SELinux MLS mode is the canonical technical implementation of Bell-LaPadula. When a question mentions SELinux with security levels, think BLP. The key: it's kernel-enforced, label-based, and no process override exists.
Biba✈️ Air Traffic Control System
β–Ό
Protects
Integrity
Key Rule
No Write Up / No Read Down
Environment
Safety-Critical Systems

An air traffic control system assigns integrity levels to data sources. Certified radar feeds from certified transponders are HIGH integrity. Unverified position reports from pilots via radio are MEDIUM integrity. Anonymous ADS-B reports from unknown sources are LOW integrity. Controllers work at MEDIUM integrity and are blocked from reading LOW integrity reports directly into their decision-making system (No Read Down). More critically, a LOW-integrity report can never overwrite a HIGH-integrity radar track (No Write Up). Lives depend on the system not being contaminated by unverified data.

🎯
CISSP Takeaway: Biba is about data trustworthiness. "Can this data be trusted enough to influence high-integrity decisions?" β€” that's Biba's question. Signals: integrity tiers + preventing contamination from low-integrity sources + safety-critical context = Biba.
Clark-Wilson🏦 Bank Double-Entry Accounting
β–Ό
Protects
Integrity
Key Rule
TP + Separation of Duties
Environment
Financial / Commercial

A bank's general ledger system implements Clark-Wilson throughout. Account balances (CDIs) cannot be edited directly β€” only via approved journal entry transactions (TPs). When a teller receives a cash deposit (UDI β€” unconstrained input), it must go through the validated deposit workflow (IVP converts UDI to CDI) before touching any account balance. Crucially, the teller who records the transaction cannot be the same person who authorizes it β€” separation of duties is enforced by the software itself. Every change is logged for audit (IVP runs continuously). The system's integrity verification (IVP) runs nightly to ensure all ledger balances are consistent.

🎯
CISSP Takeaway: Clark-Wilson was literally designed to formalize how accounting systems should work. If a scenario describes integrity enforced through PROCEDURES (not levels) + separation of duties + controlled data items, it's Clark-Wilson β€” not Biba.
Brewer-NashπŸ›οΈ Big Four Consulting Firm
β–Ό
Prevents
Conflict of Interest
Key Rule
Chinese Wall (dynamic)
Environment
Consulting / Finance

A partner at a Big Four firm begins work on a cost-reduction engagement for Coca-Cola. Their document management system automatically places all Coca-Cola files and the entire "Beverages" conflict class in their accessible zone. When PepsiCo's account team requests the same partner for a strategy project, the system flags the conflict β€” the partner has Coca-Cola data in their access history. Even the managing director cannot override this wall to assign them to PepsiCo. The partner must first be "cleaned off" the Coca-Cola engagement and wait a defined period before the wall can be lowered. This protects both clients and the firm from legal liability.

🎯
CISSP Takeaway: Two signal phrases that always mean Brewer-Nash: "conflict of interest" and "based on previously accessed data." The dynamic, growing wall that no person can override is uniquely Brewer-Nash. No other security model behaves this way.
Brewer-Nashβš–οΈ Law Firm Matter Management
β–Ό
Prevents
Conflict of Interest
Key Rule
Prior Access History
Environment
Legal Services

A law firm's matter management system runs a conflict check before assigning any attorney to a new case. Attorney Rivera has worked on patent disputes for TechCorp. When the firm is asked to represent StartupAI β€” which TechCorp is actively suing β€” the system blocks Rivera from the matter entirely. Rivera's access history has built a permanent wall between them and any party opposing TechCorp. The system even tracks former clients from previous firms if Rivera disclosed them during onboarding. This is the Chinese Wall in legal practice, required by bar association ethics rules.

🎯
CISSP Takeaway: Law firms are required by professional ethics rules to maintain these walls. The model was explicitly designed to encode these legal and ethical obligations into a formal security framework. "Chinese Wall" and "Brewer-Nash" are completely interchangeable on the exam.
PRACTICE QUIZ

10 CISSP Scenario Questions

Each question is phrased the way the CISSP exam tests security models β€” scenario first, model identification second.

Question 1 of 10
QUESTION 1
Score: 0/0
DECISION TOOL

Which Security Model Is This?

Answer 3 questions about the scenario and we'll identify the model β€” with a full explanation.

Does the scenario describe preventing the disclosure of classified or sensitive information to lower-clearance levels?
Think: Is the concern about SECRECY β€” keeping classified data from flowing down to people without clearance?
πŸ”΄
Yes β€” it's about keeping secrets secret
Clearance levels, classification labels, preventing downward information flow
➑️
No β€” it's about something other than secrecy
Integrity, procedures, or preventing conflicts of interest
Does access change dynamically based on what the user has previously accessed β€” preventing conflicts of interest?
Think: Is a "wall" being built based on prior access history? Would a user be blocked because they already accessed a competitor's data?
🟣
Yes β€” access grows more restricted over time
Consulting firm, law firm, investment bank β€” conflict of interest, Chinese Wall
➑️
No β€” it's about protecting data integrity
Preventing contamination or corruption of data
How is integrity enforced in the scenario?
Think: Is it about restricting reads/writes based on integrity LEVELS (like classification tiers), or through APPROVED PROCEDURES and separation of duties?
πŸ”΅
Through integrity levels β€” high/medium/low tiers
Low-integrity sources can't corrupt high-integrity data. No write up, no read down.
🟒
Through approved procedures + separation of duties
CDIs can only be modified via TPs. One person can't initiate and approve the same transaction.
MEMORY HOOKS

Mnemonics & Memory Tricks

Click each card to flip and reveal the mnemonic. Built for rapid recall under exam pressure.

πŸ”΄
Bell-LaPadula
Click to reveal
"No Read Up,
No Write Down"
BLP = Blocks Leaking to Public. Secrets can only go UP. They can never come DOWN.
πŸ”΅
Biba
Click to reveal
"No Write Up,
No Read Down"
Biba = Blocks Infections from Below. Dirty data stays dirty β€” can't infect clean data above.
🟒
Clark-Wilson
Click to reveal
"Controlled Workflow β€”
Procedures + SoD"
C-W = Controlled Workflow. All data changes go through approved TPs. No single person completes a full transaction.
🟣
Brewer-Nash
Click to reveal
"Build a Wall β€”
History Decides"
BN = Build a New wall based on history. Once you touch one side, you're walled off from the other. It grows automatically.

πŸ”„ BLP vs Biba β€” The Inverse Relationship

This is the single most-tested confusion point. Commit this table to memory.

Rule πŸ”΄ Bell-LaPadula πŸ”΅ Biba
Read Rule❌ No Read Up❌ No Read Down
Write Rule❌ No Write Down❌ No Write Up
ProtectsCONFIDENTIALITYINTEGRITY
Info flow directionCan only flow UP ⬆️Can only flow DOWN ⬇️

🎯 The Ultimate Security Models Exam Cheat Sheet

If the question says…Think…Model
"clearance level" / "classified" / "no read up" / "no write down" / "secrets leaking down"πŸ”΄ Military secrecy. Info flows UP only. BLP = Blocks Leaking to PublicBell-LaPadula
"integrity level" / "contamination" / "no write up" / "no read down" / "unverified source"πŸ”΅ Data purity. Dirty data stays dirty. Biba = Blocks Infections from BelowBiba
"separation of duties" / "approved procedure" / "TP" / "CDI" / "dual authorization" / "no direct edit"🟒 Accounting rules. Controlled workflow. C-W = Controlled WorkflowClark-Wilson
"conflict of interest" / "Chinese Wall" / "previously accessed" / "dynamic restriction" / "consulting/law firm"🟣 History builds the wall. BN = Build a New wall based on historyBrewer-Nash
⚠️
Clark-Wilson vs Biba trap: Both protect INTEGRITY β€” students often pick one when they mean the other. The key difference: Biba uses levels (HIGH/MEDIUM/LOW integrity tiers). Clark-Wilson uses procedures (TPs, separation of duties, CDIs). If the scenario mentions "separation of duties" or "approved transaction workflow," it's Clark-Wilson. If it mentions "integrity level" or "contamination from lower-trust source," it's Biba.
πŸŽ“ CISSP Exam Prep Platform

Ready to Pass the CISSP?
Get Everything You Need in One Place.

These infographics are just the start. FlashGenius gives you a complete CISSP prep toolkit β€” practice tests, flashcard decks, cheat sheets, and domain quizzes built for how security professionals actually learn.

🎯 Practice Tests
πŸƒ Flashcard Decks
πŸ“„ Cheat Sheets
πŸ“Š Domain Quizzes
🧠 Memory Hooks
πŸ“š Study Guides
πŸš€ Start Free on FlashGenius View All CISSP Resources β†’
Free to register Β· No credit card required Β· Trusted by CISSP candidates