Compliance requirements are obligations imposed by laws, regulations, auditors, security teams, risk officers, or enterprise customers. They merit distinct treatment from other non-functional requirements because their stakeholders, language, and consequences differ. section 2.3
Applicability is usually a yes-or-no question, not a tier: a system either falls under GDPR or processes cardholder data subject to PCI DSS, for example. Security differs: it calls for identifying concrete threats and choosing controls suited to the system’s risk. section 2.3
Bring legal, privacy, and security stakeholders into architecture discussions while options remain open. Identify constraints such as data residency, retention limits, audit requirements, and restrictions on third-party services before they make a chosen design costly to change. section 2.3.4
Choose safeguards proportionately
Protecting personal data at rest does not automatically require complex application-level encryption; storage-level cloud encryption may be reasonable. Excessive access reviews can become rubber-stamping, while blanket bans on AI agents may delay learning when regulatory guidance is absent. Weigh legal context alongside engineering costs, performance, and operational trade-offs. section 2.3.4
