Tian2
Library AP Cybersecurity Unit 2: Securing Spaces
⁂   AP Cybersecurity · Unit 2 · Weeks 8–13

2. Securing Spaces

The foundational frameworks of the entire course: CIA Triad, risk assessment, defense in depth, physical security controls, insider threats, the eight-phase attack kill chain, and the four security control types. Unit 2 generates the most exam questions of any single unit.

Weeks 8–13 Most exam questions Skill emphasis: Analyze Risk, Mitigate Risk

The CIA Triad

Every security decision in every unit maps back to one or more of the three CIA pillars. The exam presents scenarios and asks which pillar is violated, or which control protects which pillar.

PillarMeaningExample violationExample control
ConfidentialityOnly authorized parties can access the information.Employee emails a customer list to a personal account.Encryption, access controls, DLP.
IntegrityInformation is accurate and cannot be altered by unauthorized parties.Attacker modifies financial records in transit.Cryptographic hashing, digital signatures, audit logs.
AvailabilityInformation and systems are accessible when needed by authorized users.DDoS attack takes down a banking website.Redundancy, load balancing, DDoS mitigation, backup power.

Exam tip: A single attack can violate multiple pillars. Ransomware primarily violates Availability (files are locked) but may also violate Confidentiality (data is exfiltrated before encryption). When the question asks "which pillar," identify the primary violation described in the scenario.

Risk Assessment

Risk is quantified as the product of how likely a threat is and how severe its impact would be:

$$\text{Risk} = \text{Likelihood} \times \text{Impact}$$

A standard 5×5 risk matrix assigns values 1–5 to each axis. Controls are recommended in proportion to calculated risk — not every risk justifies the same investment.

Risk Assessment Process

  1. Identify assets — what must be protected? (data, hardware, processes, reputation)
  2. Identify threats — what could harm those assets?
  3. Assess Likelihood × Impact — place each threat on the matrix
  4. Recommend proportionate controls — high-risk items get layered controls; low-risk items may need only monitoring

Defense in Depth

No single control is sufficient. Layers of controls across three domains ensure that bypassing one layer does not mean a successful breach.

DomainExamples
PhysicalLocks, fences, mantraps, badge access, security cameras
Logical (Technical)Firewalls, encryption, access controls, IDS/IPS, MFA
AdministrativeSecurity policies, procedures, training, background checks, separation of duties

An attacker who tailgates into a building (bypasses a physical control) is still stopped by a login password (logical), and may be caught on CCTV reviewed after a SIEM alert (detective + administrative).

Physical Access Controls

ControlHow it worksControl type
Badge / keycardRFID or magnetic-stripe credential grants zone access. Logs entry/exit timestamps. Revocable instantly.Preventative + Detective
BiometricsFingerprint, iris, or facial recognition. Cannot be shared, stolen, or forgotten. High assurance of identity.Preventative
Mantrap / airlockTwo-door chamber: first door must close before second opens. Prevents tailgating and piggybacking.Preventative
CCTVVisible cameras deter attempts (preventative). Footage supports forensic review (detective).Preventative + Detective

Key distinction: Tailgating = following an authorized person through a door without their knowledge. Piggybacking = doing so with their (sometimes unwitting) consent. Both are defeated by mantraps; badge access alone does not stop either.

Insider Threats

An insider threat originates from someone with authorized access — employee, contractor, or business partner. Two categories:

  • Malicious insider: Intentionally abuses access for personal gain, sabotage, or espionage. Their access is legitimate, making detection harder than external threats.
  • Negligent insider: Causes harm through carelessness without malicious intent — clicking a phishing link, misconfiguring a server, losing an unencrypted laptop.

Mitigations: Least privilege (limit access to only what is needed), separation of duties (require two people for sensitive actions), UEBA (behavioral analytics to detect anomalies), exit procedures (immediate revocation of access).

The Cyberattack Kill Chain (8 Phases)

Understanding the attack sequence lets defenders identify where they can intervene to stop or limit an attack.

#PhaseWhat happensDefensive opportunity
1ReconnaissanceOSINT, network scanning, identifying targetsReduce public footprint; honeypots
2WeaponizationCrafting the exploit or malicious payloadThreat intelligence; vulnerability patching
3DeliveryEmail attachment, link, USB, or web exploitEmail filtering, user training, web proxy
4ExploitationVulnerability triggered; code executesPatch management, secure coding, sandboxing
5InstallationBackdoor, rootkit, or scheduled task installedEndpoint detection, application whitelisting
6Command & Control (C2)Compromised host beacons to attacker infrastructureDNS sinkholing, outbound firewall rules, SIEM
7ExfiltrationData stolen, files encrypted (ransomware), objective achievedDLP, network monitoring, data classification
8EvasionLogs deleted, channels encrypted, traffic blends inImmutable logs, SIEM correlation, forensics

The Four Security Control Types

TypeWhen it actsExamples
PreventativeBefore — stops the attack from occurringFirewalls, access controls, encryption, security training, mantrap
DetectiveDuring/after — identifies that an attack occurredIDS, audit logs, SIEM alerts, CCTV footage review
CorrectiveAfter — restores normal operationsIncident response plan, data backups, patch management, system reimaging
CompensatingWhen the ideal control is unavailable — equivalent protection by alternative meansManual approval workflow when automated MFA is unavailable; network segmentation to isolate an unpatched legacy system

Exam trap: A CCTV camera is both preventative (visible cameras deter) and detective (footage is reviewed after an incident). Classify by the function described in the specific scenario context.

Environmental Threats

Physical infrastructure security extends to environmental risks that directly violate Availability:

  • HVAC failure: Server rooms require continuous cooling. Redundant HVAC with automated temperature monitoring is standard data-center practice.
  • Fire: Clean-agent suppression systems (not water, which damages electronics) in server rooms. Fire detection and physical compartmentalization limit spread.
  • Power failure: UPS (uninterruptible power supply) bridges short outages; diesel generators handle extended outages. N+1 redundancy means at least one backup per critical component.
  • Natural disasters: Geographically separated secondary data centers and off-site backups protect against flood, earthquake, and severe weather scenarios.

Worked FRQ Scenario: Unit 2 Style

Original Practice Scenario · Tian2 AP

Scenario: A hospital's risk team identifies two threats: (A) a staff member accidentally emailing a patient record to the wrong address (Likelihood: High; Impact: Medium, scored 4×3=12) and (B) a nation-state attack on the medication dispensing system (Likelihood: Low; Impact: Very High, scored 2×5=10).

(A) Which threat should receive higher priority investment? Explain using the risk matrix.

Threat A (score 12) has higher priority. The risk matrix incorporates both dimensions; the accidental email threat's much higher likelihood outweighs the impact difference at these values. Risk management allocates resources proportionate to total calculated risk, not worst-case scenario alone.

(B) For Threat A, identify the CIA pillar violated and recommend one preventative and one detective control.

Pillar: Confidentiality — patient data is disclosed to an unauthorized party.

Preventative: Data Loss Prevention (DLP) software scans outbound email for patterns matching protected health information (PHI) and blocks or quarantines the message before delivery.

Detective: Email audit logging with automated SIEM alerts when messages containing large volumes of structured data are sent to external addresses outside an approved whitelist.