Security By Design
summary
Argues that security must be baked into architecture by mapping business risk tolerance to features, and shows concrete patterns that work: small functional APIs, multiparty authorization, zero trust/BeyondCorp with precomputed device trust, SLOs/error budgets for reliability, auditing and break‑glass recovery. Illustrates failures: a 2012 password‑manager outage after thousands of simultaneous requests (five years uptime beforehand) stalled recovery for hours because an HSM card was inserted backwards. Recommends specific controls, telemetry and partitioning (multiple SLOs for Bigtable workloads) and when to accept tradeoffs.
AI-written summary · about
chapters (9)
- 0:00Intro
- 2:22Agenda
- 3:13Security & reliability
- 9:49Managing security risk
- 15:42Managing reliability risk
- 19:25Deep dive: Design for least privilege
- 37:56Design strategies to consider
- 40:23Shift security & reliability left
- 41:49Outro
related talks
01h 17mSecure by Design - Security Principles for the Working Architect
50mSecure and Privacy by Design Converge with Threat Modeling
34mSecure by design - the secret weapon of high-performing organisations
45mSecurity Architecture, An Adventure?
51mSecure Applications, By Design
16mDesigning Holistic Security Programs
28mSecurity Through Design - Making Security Better By Designing for People
18mDesigning for Security Outcomes