[reference] — from the roadmap.sh cyber-security roadmap. How you build systems to be defensible in the first place — the strategy that everything else in security operations operates within.
Defense in depth
The foundational principle: no single control is trusted to stop everything, so you layer them. A firewall and segmentation and endpoint protection and MFA and logging and least privilege — so a failure or bypass of one layer is caught by the next. An attacker must defeat every layer; a defender only needs one to catch them.
perimeter (firewall) → network (segmentation) → host (hardening/EDR)
→ application (secure coding) → data (encryption) → identity (MFA, least privilege)
...with monitoring across all layers
This connects to the CIA triad: each layer protects confidentiality, integrity, and availability from a different angle.
Zero Trust
The modern evolution beyond “castle-and-moat” (trust everything inside the perimeter). Zero Trust assumes the network is already hostile and the perimeter is gone (cloud, remote work, mobile): never trust, always verify. Its tenets:
- Verify explicitly — authenticate and authorize every request based on identity, device health, and context — not on network location. Being “inside the network” grants nothing.
- Least privilege — give the minimum access needed, just-in-time; assume any account can be compromised so limit its blast radius.
- Assume breach — design as if attackers are already inside: microsegment, encrypt everywhere, monitor continuously.
Zero Trust is why a phished credential (social engineering) or a foothold on one host doesn’t automatically reach everything — it’s the architectural answer to lateral movement.
Hardening
System hardening shrinks the attack surface by removing what isn’t needed and locking down what remains:
- Disable unused services, ports, and default accounts; change default credentials.
- Apply the principle of least functionality — a server runs only what it must.
- Enforce secure configuration baselines (the CIS Benchmarks — frameworks).
- Patch promptly, enforce least-privilege permissions, enable OS security features (ASLR/DEP, SELinux/AppArmor, Linux controls).
- Group Policy (Windows) / configuration management (Ansible — IaC) to enforce hardening at scale.
Every disabled service is one fewer thing to exploit or patch — hardening directly reduces the exploit surface.
The teams: blue, red, purple
Security operations is organized around adversarial roles:
- Red team — offensive; simulates real attackers to find weaknesses (ethical hacking / pentesting is the discipline).
- Blue team — defensive; monitors, detects, and responds (the SOC, this whole section).
- Purple team — collaboration between the two: red attacks, blue detects, and they share findings to improve detection in a tight loop. The most productive model — offense and defense sharpening each other rather than working in silos.
Why architecture comes first
Detection, response, and forensics (the rest of this section) are what you do when prevention fails — and it will fail. But a well-architected system (defense-in-depth + zero-trust + hardening) means failures are contained, noisy, and survivable, rather than catastrophic. You can’t monitor your way out of a flat, unhardened, over-permissioned network — architecture determines how much the blue team can actually defend.
Related
- Logging, SIEM & Detection — monitoring the layers this builds
- Network Segmentation — a core defense-in-depth/zero-trust layer
- Governance, Risk & Compliance — the frameworks that mandate this
Software engineering
Infrastructure & security
Computer science
Hardware & engineering