Konstas.Engineering Notes
Networking · Cybersecurity · Infrastructure
Security

Information Security Fundamentals: CIA, Risk, Access Control, Cryptography & Compliance

Articles · 2026 · Ioannis (John) Konstas

Sound information security practice rests on a small number of foundational concepts that appear repeatedly across every framework, standard, and regulation in the field. Understanding these concepts deeply — not just their definitions, but their relationships and practical implications — is what separates a security practitioner from someone who has memorised a glossary.

This article covers: the CIA triad, identity and access frameworks, risk quantification, security controls taxonomy, access control models, cryptographic fundamentals, key compliance obligations, and physical and network security principles.


The CIA Triad

PillarDefinition
ConfidentialityEnsures that information is disclosed only to authorised individuals, programs, and processes. Controls include encryption, logical access controls, physical access restrictions, and need-to-know enforcement. The opposite threat is unauthorised Disclosure.
IntegrityGuarantees that data is not modified in an unauthorised or undetected manner. Data must be consistent and accurate throughout its lifecycle. The opposite threat is Alteration — whether accidental or malicious.
AvailabilityEnsures that systems and data are accessible and usable when needed by authorised parties. Achieved through fault tolerance, redundancy, backup, and recovery procedures. The opposite threat is Destruction or denial of service.

Identity, Authentication, Authorisation, and Accountability (IAAA)

TermDefinition
IdentificationThe act of a subject claiming an identity — a username, employee ID, or certificate. Used as the basis for all subsequent access control decisions.
AuthenticationTesting the evidence of identity. Proves the subject is who they claim to be through something they know (password), something they have (token), or something they are (biometric).
AuthorisationDefining and enforcing what rights and permissions an authenticated identity is granted — what resources they may access and what actions they may perform on them.
AccountabilityThe ability to trace specific actions to a specific individual. Requires that actions are logged and that those logs are protected from modification. Non-repudiation is the evidence dimension of accountability.

Risk Quantification

FormulaLabelDescription
SLE = Asset Value × Exposure FactorSingle Loss ExpectancyThe dollar value lost when an asset is successfully attacked. The exposure factor (0–1) represents the percentage of the asset's value destroyed in a single incident. If a server worth $100,000 is 40% likely to be destroyed in a fire, SLE = $40,000.
ALE = SLE × AROAnnual Loss ExpectancyThe expected annual cost of a threat. ARO is the Annualised Rate of Occurrence — how frequently the event is expected in a year. A once-per-century event has ARO = 0.01. ALE drives control investment decisions: a control costing more than ALE is economically unjustifiable.
Residual Risk = Total Risk − Controls GapResidual RiskThe risk remaining after controls are implemented. The controls gap is the reduction in risk achieved by the safeguards in place. Organisations do not eliminate risk — they reduce it to an acceptable level. Legally, remaining residual risk is not held against an organisation when due care has been exercised.

Risk Responses

ResponseDefinition
AvoidanceDiscontinue the business activity that generates the risk. Complete elimination of exposure but also of the associated business benefit.
TransferPass the financial consequences of the risk to another entity — typically through insurance or contractual indemnification. The risk event itself is not eliminated, only the financial exposure.
MitigationImplement controls to reduce the likelihood or impact of the risk to an acceptable level. The most common response. Cost of mitigation should not exceed ALE.
AcceptanceAcknowledge the risk and consciously decide to live with it. Appropriate when the cost of mitigation exceeds the potential loss, or when residual risk after mitigation is below the organisation's risk threshold.

Security Controls Taxonomy

By Type

TypeExamplesPreventiveDetective
Administrative / ManagerialHiring policies, security awareness training, background checks, separation of duties, mandatory vacations, rotation of duties, acceptable use agreementsPolicies, screening, awareness trainingJob rotation, audit record review, background verification
Technical / LogicalEncryption, firewalls, IDS/IPS, access control lists, biometrics, smart cards, audit logs, PKIEncryption, firewalls, authentication mechanismsIDS, automated violation reports, audit trail analysis
PhysicalFences, locks, guards, doors, windows, CCTV, motion detectors, badge readersFences, locks, security guards, badge accessCCTV, motion sensors, thermal detectors

By Category

CategoryDefinition
DirectiveSpecify rules of behaviour — acceptable use policies, codes of conduct, legal compliance requirements.
DeterrentDiscourage potential attackers by raising the perceived cost or risk of an attack — warning banners, visible cameras, legal notices.
PreventiveStop an incident from occurring — access controls, encryption, firewalls, physical locks.
CompensatingAlternative controls that substitute for a primary control when it cannot be implemented — used when a direct control is cost-prohibitive or technically infeasible.
DetectiveSignal that an incident has occurred or is occurring — IDS alerts, audit logs, CCTV recording.
CorrectiveMitigate damage and restore systems after an incident — patch deployment, incident response procedures.
RecoveryRestore full operational capability after an incident — backup restoration, disaster recovery failover.

Administrative Controls

ControlDefinition
Separation of dutiesAssigns different parts of a sensitive task to different individuals so no single person has end-to-end control. Prevents both accidental errors and deliberate fraud. Effective against insider threats and collusion.
Least privilegeUsers receive only the minimum rights necessary to perform their assigned function, and only for the duration required. Three access levels: read-only, read/write, and access/change. Reduces the blast radius of a compromised account.
Two-man / Dual controlRequires two individuals to complete a sensitive operation. Two-man control means both persons review each other's work. Dual control means both are required to complete the action at all. Used for key escrow recovery, critical configuration changes, and financial authorisations.
M of N ControlA generalisation of dual control: a minimum of M agents out of N total authorised agents must cooperate to perform a high-security operation. Example: 3 of 8 key escrow recovery agents must act together to recover a single key.
Rotation of dutiesLimits how long a person holds a specific security-sensitive role. Reduces collusion risk and allows independent review of prior decisions. Mandatory vacations serve a similar purpose — fraudulent schemes often require continuous presence to maintain.
Need to knowA subject is given only the specific information required to complete their assigned task — even if their clearance level would permit broader access. Clearance is necessary but not sufficient; need to know is an additional gate.

Access Control Models

ModelFocusOriginNotes
Bell-LaPadulaConfidentiality (MAC)U.S. Department of DefenseDesigned to prevent information from flowing from high to low classification. Addresses confidentiality only — not integrity.
BibaIntegrity (MAC)Inverse of Bell-LaPadulaDesigned to prevent information from flowing from low integrity to high integrity — protecting data from contamination by untrusted sources. Addresses integrity only, not confidentiality.
Clark-WilsonIntegrity (commercial)Commercial environmentsThe first model designed for commercial use. Enforces well-formed transactions and separation of duty — more practical than Bell-LaPadula or Biba for enterprise environments.
Brewer-Nash (Chinese Wall)Conflict of InterestFinancial / consulting environmentsPrevents members of an organisation from accessing information that creates a conflict of interest with data they have already viewed. Relevant in financial services, legal, and consulting contexts.

Cryptography

Symmetric Algorithms

AlgorithmDetail
DES56-bit key, 64-bit blocks, 16 rounds of substitution/transposition. Replaced by AES. TripleDES (3DES) applies DES three times with up to three keys (168-bit key length; effective security is approximately 112 bits due to the meet-in-the-middle attack, 48 rounds).
AES (Rijndael)128/192/256-bit keys, 128-bit blocks. NIST standard since 2001. Used in BitLocker, Microsoft EFS, WPA2, and US government classified data — AES-128 is approved for Secret, AES-256 for Top Secret. The current symmetric encryption standard.
Blowfish32–448 bit variable key length. Used in bcrypt on Linux systems. Alternative to DES on resource-constrained systems. By Bruce Schneier.
RC4Stream cipher. Used in WEP and early WPA (TKIP). WEP is broken and should never be used. RC4 itself is considered weak.

Asymmetric Algorithms

AlgorithmDetail
RSABased on the difficulty of factoring the product of two large prime numbers (trapdoor function). Supports encryption, key exchange, and digital signatures. Standard for asymmetric encryption.
Diffie-HellmanKey exchange protocol — not for encryption directly. Enables two parties to establish a shared secret over an insecure channel without ever transmitting the secret itself.
ECCElliptic Curve Cryptography. Equivalent security to RSA at much smaller key sizes. Requires fewer computational resources — standard for mobile and embedded systems.
DSADigital Signature Algorithm. The US government's standard for digital signatures. Used in conjunction with SHA for signing — not for encryption.

Hash Algorithms

AlgorithmDetail
MD5128-bit digest. Produces output in four rounds over 512-bit blocks. Subject to collision attacks — two different certificates can produce the same MD5 hash. Not suitable for integrity verification of high-value data or password storage.
SHA-1160-bit digest. Designed by NIST/NSA. No longer considered secure for new systems — collisions have been demonstrated.
SHA-2 / SHA-3Current standards. SHA-256 and SHA-512 are widely deployed. SHA-3 is the newest standard though SHA-2 remains the dominant production choice.

Key Laws and Regulations

LawAreaDetail
SOX (Sarbanes-Oxley, 2002)Financial / CorporateSection 302: CEOs and CFOs personally certify financial statements — criminal liability for knowingly signing incorrect information. Section 404: Internal controls assessment over financial reporting. COSO is the supporting framework for SOX 404 compliance.
HIPAA (1996) + HITECH (2009)HealthcareHIPAA establishes privacy and security requirements for Protected Health Information (PHI). HITECH extended HIPAA obligations directly to Business Associates (BAs) and introduced mandatory breach notification requirements. Any covered entity–BA relationship must be governed by a Business Associate Agreement (BAA).
GLBA (Gramm-Leach-Bliley Act)Financial / PIIRequires financial institutions to protect the privacy of customer non-public personal information. Mandates a written information security program, privacy notices, and third-party service provider oversight.
FERPAEducationProtects the privacy of student educational records. Requires parental or student consent before records are disclosed to third parties.
CFAA (Computer Fraud and Abuse Act, 1986/1996)Computer CrimeFederal computer crime law. Criminalises unauthorised access to computers, trafficking in computer passwords, and attacks that cause losses of $1,000 or more or could impair medical treatment.
ECPA (Electronic Communications Privacy Act, 1986)Communications PrivacyProhibits eavesdropping or interception of electronic communications. CALEA (1994) amended ECPA to require communications carriers to support lawful wiretap capability regardless of technology.
FISMA (2002)Federal GovernmentRequires federal agencies to implement an information security programme covering categorisation (Phase 1), selection of minimum controls, and assessment. NIST SP 800 series provides the implementation framework.
GDPR (General Data Protection Regulation, 2018)Privacy (EU)Seven principles (Article 5): Lawfulness/Fairness/Transparency, Purpose Limitation, Data Minimisation, Accuracy, Storage Limitation, Integrity/Confidentiality, and Accountability. Individuals have the right to access, correct, and erase their data. Data must be collected for a specified, explicit, and legitimate purpose and not further processed in an incompatible manner. Requires explicit consent or a lawful basis for processing.

Physical Security: Fire Classification

ClassFuelSuppression AgentNote
ACommon combustibles (wood, paper)Water, Soda AcidWater suppresses temperature
BLiquid and gas fuelsCO2, Dry ChemicalCO2 removes oxygen; dry chemical interrupts the combustion chain. Never use water or soda acid on Class B fires.
CElectrical equipmentCO2, non-conductive gasNever use water on electrical fires
DCombustible metalsDry powderSpecialised agent only

Physical Security: Power Events

EventTypeDurationResponse
SpikeExcessShortSurge protector
SurgeExcessProlongedSurge protector / UPS
FaultLossMomentaryUPS
BlackoutLossExtendedGenerator + UPS
Sag / DipReductionMomentaryUPS / CVT
BrownoutReductionProlongedConstant Voltage Transformer

Network Security: Firewall Generations

GenerationNameLayerDetail
1st GenStatic Packet FilteringLayer 3–4Examines source/destination IP, protocol, and port. Uses ACLs to permit or deny. Fast but cannot inspect session state or payload content. Easily evaded by fragmentation attacks.
2nd GenStateful Packet InspectionLayer 4–7Maintains a state table of established connections. Makes decisions based on conversation context, not just individual packets. Detects anomalies that stateless filters miss — such as SYN floods and unexpected packet sequences.
3rd GenApplication-Level ProxyLayer 7Operates as a relay — client connects to proxy, proxy connects to server. Inspects full payload, can authenticate, re-encrypt, and log application-layer content. Slowest but most thorough inspection.
CircuitCircuit-Level ProxyLayer 5Monitors TCP/UDP handshaking to verify session legitimacy. Once the circuit is established, data tunnels between parties without further inspection. Less granular than application proxies but covers a wider protocol range.

Intellectual Property

TypeDurationDetail
Patent20 years from applicationGrants the owner exclusive rights to an invention. After 20 years the invention enters the public domain. Requires public disclosure (registration) — trade-off: protection in exchange for disclosure.
Copyright70 years after author's deathProtects the expression of ideas, not the ideas themselves. Applies automatically upon creation — no registration required. Covers software, documentation, creative works.
TrademarkRenewable every 10 yearsProtects words, names, symbols, or combinations used to identify a product or service. Distinguishes the source of goods from competitors.
Trade SecretIndefinite (while kept secret)Protects proprietary business information critical to commercial advantage — formulas, processes, customer lists. No registration. Protection is lost if the secret is disclosed or independently discovered.

← All articles

Research and direction by Ioannis (John) Konstas; compiled and finalized with AI assistance.

👁· views