Multilevel and Multilateral Security
Stub. Access control models richer than a basic ACL/RBAC: multilevel security (mandatory, lattice-ordered classification levels, e.g. Bell-LaPadula's "no read up, no write down") and multilateral security (need-to-know compartmentation across peer groups rather than a strict hierarchy, e.g. Chinese-wall policies).
Basic access control (an ACL per object, or a role per subject) assumes permissions form a simple, discretionary relation. Multilevel security formalizes a stricter, mandatory alternative used in classified/military settings: information has a classification level, subjects have a clearance level, and information flow is constrained by a lattice order (Bell-LaPadula’s confidentiality model: no reading above your clearance, no writing below it, to prevent high-classification information leaking down). Multilateral security instead compartmentalizes horizontally — need-to-know groups at the same sensitivity level that still shouldn’t see each other’s information (Chinese-wall policies, common in professional-services conflict-of-interest settings) — a genuinely different constraint that a pure hierarchy can’t express.
Outline (TODO — flesh out each)
- Multilevel security — classification lattices, Bell-LaPadula
- “No read up, no write down”
- Multilateral security — compartmentation, need-to-know
- Chinese-wall policies