(ISC)² CISSP Security Architecture and Engineering Practice Questions

Free CISSP Security Architecture and Engineering practice test questions

8 questions from this domain with answers and explanations - different from the samples on the main (ISC)² CISSP page. Sign up free to practice the full set.

1. Question 1Security Architecture and Engineering

   What is 'defense in depth'?

   • AEnforcing one strong authentication mechanism as the sole layer of the whole defense strategy
   • BRelying on a single hardened perimeter control placed only at the network boundary
   • CApplying encryption only to backups while leaving every other security layer wide open
   • DLayering multiple, diverse controls so a single failure doesn't lead to compromiseCorrect
   ✓ Correct answer: D

   Defense-in-depth uses multiple layers of security controls; if one control fails, others continue to protect assets. Layers include perimeter controls (firewalls), network controls (IDS/IPS), host controls (antivirus), application controls (input validation), and data controls (encryption). Defense-in-depth assumes that no single control is 100% effective. Redundant controls for critical assets ensure protection despite single control failures. Each layer should be independent so compromise of one doesn't disable others.

   Why the other options are wrong

   • AA single mechanism, however strong, is one point of failure rather than layered defense.
   • BA hard shell with a soft interior fails the moment the perimeter is breached.
   • CProtecting only backups leaves most attack paths unlayered and undefended.
2. Question 2Security Architecture and Engineering

   Which AES mode should be avoided because identical plaintext blocks produce identical ciphertext blocks, leaking patterns?

   • AECB (Electronic Codebook)Correct
   • BCBC (Cipher Block Chaining)
   • CCTR (Counter)
   • DGCM (Galois/Counter Mode)
   ✓ Correct answer: A

   ECB encrypts each block independently with the same key and no chaining or nonce, so identical plaintext blocks always produce identical ciphertext blocks, leaking structural patterns; this is why ECB is discouraged beyond a single block. CBC, CTR, and GCM all mix in a prior ciphertext block, counter, or nonce, so identical plaintext no longer yields identical ciphertext.

   Why the other options are wrong

   • BCBC chains each block with the previous ciphertext block, so identical plaintext blocks produce different ciphertext.
   • CCTR encrypts a unique counter value per block, so repeated plaintext does not repeat in the ciphertext.
   • DGCM combines counter-mode encryption with authentication, avoiding ECB's pattern leakage and adding integrity.
3. Question 3Security Architecture and Engineering

   A team stores user passwords using a single round of unsalted SHA-256 and asks why this is risky after a database leak. What is the BEST explanation and remedy?

   • AFast unsalted hashes are vulnerable to rainbow tables and GPU cracking; use a slow salted KDFCorrect
   • BSHA-256 is the weakness; switching to SHA-512 lengthens the digest and stops cracking
   • CAdd a single global pepper stored beside the hashes so every digest becomes unique
   • DApply many thousands of extra SHA-256 iterations, which alone makes digests uncrackable
   ✓ Correct answer: A

   SHA-256 is a fast, general-purpose hash designed for speed, so without a per-user salt, identical passwords produce identical digests that attackers can crack in bulk using precomputed rainbow tables or brute-force GPU hardware. A slow key-derivation function like bcrypt, scrypt, or Argon2 combines a unique salt with deliberate computational cost, making large-scale offline cracking after a leak impractical even though the hashes are exposed.

   Why the other options are wrong

   • BSHA-512 is still a fast, unsalted general-purpose hash, so lengthening the digest does not stop GPU cracking or rainbow tables.
   • CA pepper stored next to the hashes is disclosed in the same leak, and a global pepper does not make per-user digests unique.
   • DIterating a fast hash without a per-user salt still leaves all users exposed to shared rainbow-table precomputation.
4. Question 4Security Architecture and Engineering

   Anticipating future quantum computers, an organization wants to protect today's encrypted traffic against a 'harvest now, decrypt later' adversary. Which strategy most directly addresses this concern?

   • AAdopt post-quantum key-establishment algorithms now, optionally hybridized with classical algorithmsCorrect
   • BIncrease symmetric AES keys from 128 to 256 bits and treat the harvested traffic as adequately protected
   • CMigrate all RSA key exchange to 4096-bit keys to stay ahead of anticipated quantum attacks
   • DRely on ephemeral perfect forward secrecy alone to keep the captured ciphertext permanently undecryptable
   ✓ Correct answer: A

   A 'harvest now, decrypt later' adversary records ciphertext today expecting a future quantum computer to break the asymmetric key exchange via Shor's algorithm. Deploying NIST-selected post-quantum key-encapsulation mechanisms, frequently in a hybrid mode with classical ECDH for safety, protects the long-term confidentiality of data exchanged now. This addresses the asymmetric components Shor's algorithm threatens most.

   Why the other options are wrong

   • BLarger AES keys help symmetric strength but the harvested traffic's key exchange still relies on quantum-vulnerable public-key algorithms.
   • CLonger RSA keys are still broken by Shor's algorithm on a quantum computer, so they do not protect harvested traffic.
   • DForward secrecy still uses quantum-vulnerable key agreement, so a future quantum attacker can recover the session keys from captured handshakes.
5. Question 5Security Architecture and EngineeringSelect all that apply

   Which TWO design choices best help prevent insecure direct object access and ensure proper isolation in a multi-tenant SaaS architecture? (Choose TWO)

   • AEnforce server-side authorization checks that scope every data request to the authenticated tenantCorrect
   • BRely on unguessable random object identifiers as the sole access control
   • CInclude a tenant identifier in queries and validate it against the session on the serverCorrect
   • DPerform tenant-isolation checks only in the client-side JavaScript
   ✓ Correct answer: A, C

   Insecure direct object references and tenant bleed are prevented when the server independently confirms that the requesting tenant is authorized for each object, rather than trusting identifiers supplied by the client. Binding every query to the session's tenant and validating that scope server-side ensures one tenant cannot reach another's data. These checks must occur on the trusted server tier.

   Why the other options are wrong

   • BUnguessable identifiers add obscurity but are not access control; a leaked or logged ID still grants access without authorization checks.
   • DClient-side checks can be bypassed by manipulating requests directly, so they cannot enforce isolation.
6. Question 6Security Architecture and Engineering

   A web application stores user passwords as unsalted SHA-256 hashes. After a breach, attackers crack many accounts quickly using precomputed tables. Which combination of changes BEST addresses the weakness that enabled the fast cracking?

   • AAdd a unique salt per password and use a slow, tunable KDF such as bcrypt or Argon2Correct
   • BSwitch from SHA-256 to SHA-512 with no salt because a longer hash defeats precomputed tables
   • CEncrypt the password hashes with AES using a single application-wide key
   • DHash each password twice with SHA-256 to double the computational cost for attackers
   ✓ Correct answer: A

   Two separate weaknesses enabled the fast cracking: the absence of a salt let attackers reuse precomputed rainbow tables across every account, and SHA-256's speed as a general-purpose hash let brute-force guessing run at enormous rates. A unique random salt per password defeats precomputation entirely, while a deliberately slow, tunable KDF such as bcrypt, scrypt, or Argon2 raises the cost of each individual guess, so the fix must address both the missing salt and the wrong algorithm choice together.

   Why the other options are wrong

   • BA longer fast hash without a salt is still open to precomputed-table attacks.
   • CEncrypting hashes under one shared key adds a single point of failure, not salting or slowness.
   • DHashing twice with a fast algorithm barely raises cost and adds no salt.
7. Question 7Security Architecture and Engineering

   An architect is designing a system so that a single compromised component or a single malicious insider cannot complete a high-value action alone. Issuing a payment requires one employee to create it and a different employee to approve it. Which secure design principle is being applied, and how does it differ from least privilege?

   • ASeparation of duties, which divides a sensitive task among multiple people; least privilege limits each person's access to only what their role needsCorrect
   • BLeast privilege, which is the principle that requires two separate approvers; separation of duties merely limits each person's permissions across the system
   • CDefense in depth, which is the principle that requires two separate approvers; separation of duties is only about layering technical controls across the network
   • DFail-secure, which is the principle that requires two separate approvers; least privilege ensures the system denies access on any failure by default
   ✓ Correct answer: A

   Separation of duties requires that a critical, high-value process be divided so that completing it needs more than one person - here a creator and a separate approver - which thwarts both single-insider fraud and single-component compromise. Least privilege is distinct: it limits each individual subject's permissions to only what their role requires. They are complementary but address different risks, and the scenario describes separation of duties.

   Why the other options are wrong

   • BThis reverses the definitions; requiring two distinct people for one task is separation of duties, not least privilege.
   • CDefense in depth is about layered controls, not dividing a single task between two people.
   • DFail-secure concerns the safe state on failure, not splitting a task across approvers.
8. Question 8Security Architecture and Engineering

   Two parties already share a symmetric secret and only need to verify that messages between them were not altered in transit, without any requirement to prove authorship to a third party. Which mechanism is the MOST appropriate and efficient choice?

   • AAn HMAC using the shared secretCorrect
   • BAn RSA digital signature
   • CA plain SHA-256 hash of the message
   • DAn ECDSA digital signature
   ✓ Correct answer: A

   An HMAC provides integrity and origin authentication between parties who share a symmetric key, and it is far cheaper than public-key operations. Because there is no need for third-party non-repudiation, a symmetric MAC is the efficient fit. Digital signatures add non-repudiation via private keys but cost more and are unnecessary here; a bare hash provides no authentication.

   Why the other options are wrong

   • BRSA signatures add non-repudiation the scenario does not need and cost far more.
   • CA bare hash gives integrity only if the hash itself cannot be swapped by an attacker.
   • DECDSA also provides non-repudiation via keys, which exceeds this requirement.

How Security Architecture and Engineering is tested

This domain holds 141 of the 1,117 questions in the CISSP bank, about 13%. The mix is 120 single-answer multiple choice and 21 multiple-response, so it is worth practising the formats as well as the content.

Once you have a few attempts recorded, CertGrid scores every domain separately and points you at the weakest one, so you can drill Security Architecture and Engineering on its own rather than re-running full-length mocks.

Other CISSP exam domains

• Security and Risk Management172 questions
• Asset Security109 questions
• Communication and Network Security138 questions
• Identity and Access Management138 questions
• Security Assessment and Testing130 questions
• Security Operations157 questions
• Software Development Security132 questions
• All CISSP practice questions1,117 total
• Security Architecture and Engineering study notesKey concepts
• Security practice examsAll Security

CISSP Security Architecture and Engineering FAQ