A risk is a possible adverse consequence of a decision. In engineering, risk includes what might go wrong with a system, but also what a proposed safeguard or technology choice might cost the organisation. Assess the consequence alongside its business and operational context rather than assuming the most restrictive option is always safest. section 2.3.4 section 4.4.4
Two decisions call for different risk questions:
- For compliance controls, weigh data sensitivity and legal obligations against implementation cost and operational burden. Storage-level cloud encryption may meet a requirement without a complex application-level scheme; repeated access approvals may instead become rubber-stamping. section 2.3.4
- For external products, look beyond current technical quality, support contracts, and roadmaps. Ask whether the vendor has incentives to keep investing, and whether a business failure or acquisition could force a replacement. section 4.4.4
A capable product can still disappear
Layer attracted investment and customers but shut down in 2019 after failing to establish a sustainable business, leaving dependent teams to replace part of their architecture. After Qlik acquired Talend in 2023, Talend Open Studio was retired in 2024. Neither product quality nor a change of ownership alone guarantees continuity. section 4.4.4
Make the reasoning behind a control explicit. Requirements, regulations, threat models, and teams can change; without a recorded link between a requirement and its control, future engineers may not know why a log-retention period or API rate limit exists. section 2.3.4












