CISSP 8.2 - Techniques for Ensuring CIA

This episode of the ISC2 Certified Information Systems Security Professional (CISSP) exam prep series looks at the low-level techniques from Domain 3 that keep programs honest — the quiet mechanics that box a running process in, keep its failures contained, and decide who may touch what, plus the exam’s favorite distinction between trusting a mechanism and staying confident that it still works.

Frequently Asked Questions

What is confinement, and how does it box a process in?

Confinement restricts what a process is allowed to do, letting it read from and write to only the memory and resources it was granted, a technique also called sandboxing. It is the principle of least privilege applied to a running program, and its goal is to stop data from leaking to places it should never reach. If a process reaches beyond its authority, the request is denied, the attempt may be logged, and the offender is often shut down.

What are bounds, and who enforces them?

Bounds are the specific limits on which memory addresses and resources a process may touch. Every process gets an authority level, often as simple as ordinary user or privileged kernel, and that level sets its bounds. In most systems memory is carved into logical segments enforced by the operating system, while stricter systems use physical bounds that give each process its own physically separate region of memory, which is more secure but also more expensive.

How does isolation stop one crash from spreading?

Isolation is what you get when bounds are actually enforced: an isolated process can only affect its own memory and resources, so its behavior cannot reach out and disturb anything else. This protects the operating system kernel and every other application from a single misbehaving program, and it enables a fail-soft environment where one process can crash without dragging the others down.

How do access controls decide who touches what?

Access controls limit which objects a subject may reach and what it may do with them. Access rules spell out which objects are valid for each subject, and even then an object might allow one kind of action while forbidding another, such as letting someone read a file but not delete it. There are several flavors, including discretionary, role-based, and mandatory access control.

What is the difference between trust and assurance?

Trust is the presence of a security mechanism: a trusted system has protection features that work together to handle sensitive data while staying stable and secure. Assurance is the level of confidence that those features actually deliver the security they promise. Assurance is not permanent, because a vendor patch, a fresh exploit, or simply the passage of time can erode it, so it must be continually reverified through change, patch, and configuration management.

Reference: This article is based on concepts discussed in CISSP 8.2 - Techniques for Ensuring CIA.