Articles

Segmentation and trust domains: what a boundary has to enforce

A VLAN draws the segment; only an enforcement point decides what may cross it.

Reading: 5 minNetworking

Article cover: Segmentation and trust domains: what a boundary has to enforce

A trust domain is the set of devices that share one security requirement, and so one policy. Segmentation draws that set; the boundary is where the policy is applied to the traffic crossing it. They are different objects, usually on different devices, and the failure is to treat them as one — because two virtual LANs (VLANs) cannot see each other’s broadcasts, an operator assumes a policy separates them. No VLAN blocks an inter-VLAN flow on its own; something has to read the packet and decide.

The model: a domain, and the point that enforces it

   trust domain                 boundary                    trust domain
   hosts that share       the point where the policy      hosts that share
   one security            is applied, per flow           another security
   requirement                                            requirement
        |                                                       |
        +------ traffic ----->  [ read · permit · deny ]  ----->  +
                                      |
                     everything beyond the boundary shares
                     one level of trust: whatever this point
                     could not tell apart

The domain is defined by the requirement its members answer to, not by where they sit; the boundary by the policy it can apply, not by the cable it stands on. Everything beyond the boundary shares one level of trust: whatever it could not tell apart.

A VLAN separates traffic; a boundary enforces policy

SP 800-82r3 gives the VLAN its exact scope: “A VLAN can be used to logically separate areas within a network when physical separation may not be feasible due to cost or other prohibitive measures” (nist-sp800-82r3.txt lines 11049–11050), each port “only logically connected to the other ports within its VLAN” (lines 11059–11060). That is separation of broadcast domains, nothing more. Enforcement is a separate job on separate devices — gateways and firewalls, “only permitting explicitly authorized communication between segments” (lines 5276–5277) — and SP 800-215 draws the same line: segments come from network layer abstractions “such as VLAN IDs”, but “Policies apply to subnets (e.g., VLANs) and not to individual hosts” (nist-sp800-215.txt lines 1193–1195). The enforcement point sits where the domains meet: a routed subinterface carrying an access control list (ACL), a perimeter firewall that keeps state, the filtering function of the switch itself — placed, as SP 800-41r1 puts it, at “the edge of logical network boundaries” (lines 942–943). The VLAN is the segment; the enforcement point is the policy: segments without one are a domain on paper. Their mechanics belong to the sibling sheets on ACLs and firewall policy.

What the boundary cannot distinguish

SP 800-207 names the remainder: the implicit trust zone “represents an area where all the entities are trusted to at least the level of the last PDP/PEP gateway” (nist-sp800-207.txt lines 483–484). What the boundary could not separate is one domain by default. SP 800-82r3 states the consequence plainly: boundary protection devices between zones “do not mitigate lateral movement risks within a zone” (lines 4048–4049), because everything inside the perimeter is treated as trusted (lines 4046–4047) — hence “the implicit trust zone must be as small as possible” (nist-sp800-207.txt lines 492–493).

Choosing the domains: requirement, not convenience

SP 800-82r3 lists the criteria: “devices may be segmented based on management authority, level of trust, functional criticality, data flow, location, or other logical combinations” (lines 3879–3880). Function and criticality lead; physical convenience does not. SP 800-41r1 applies the same rule: a firewall belongs where “security requirements vary among their internal networks” (lines 934–937). Ignore the criterion and the domain flattens — the weakest host that shares the policy sets it for every other — as SP 800-215 records, where everything inside the firewall was “totally trusted” until lateral movement “annulled the concept of that perimeter and greatly expanded the attack surface” (lines 492–494).

Zones, conduits, and the smaller domain

The OT vocabulary names the same two objects: a zone is the group of assets answering to one requirement, a conduit the path their communications take between zones. SP 800-82r3 carries the same two objects and states the order of work — grouping components “into levels, tiers, or zones is a precursor activity” before the isolation devices that protect the communications between them (lines 3898–3902); the zone is a label on the group, not a substitute for the device (line 9324). Micro-segmentation is the same principle at the smallest useful scale — “a degree of isolation to prevent attack escalation” (nist-sp800-215.txt lines 1130–1131) — with the boundary moving to the hypervisor, a host agent or the application’s own identity, bringing the enforcement point closer to the resource (nist-sp800-207.txt lines 495–496). Zone membership and conduit exceptions are organisational decisions for Cybersecurity Governance.

Limits and the common mistake

The mistake runs both ways. One is reading a VLAN as a policy — treating “the segments exist” as “the policy exists”, when nothing between the segments reads a packet. The other is stacking boundaries faster than the policy model, so an inner device assumes the outer one blocks a flow; SP 800-41r1 warns this “requires good planning and policy coordination to prevent inadvertent security lapses” (lines 1082–1083). A boundary can enforce only what it can read — “Firewalls can only work effectively on traffic that they can inspect” (line 893) — and each one costs something to run.

What to remember

  • A trust domain shares one requirement and therefore one policy.
  • A VLAN separates traffic; a boundary enforces policy — the segment is not the policy.
  • Whatever a boundary cannot distinguish is one domain: keep it small.
  • Micro-segmentation: same principle, smaller domain, boundary closer to the resource.

Level and prerequisites

L3 — integration and technical control: the sheet coordinates concepts met separately at L1 and L2 (broadcast domains, VLANs, ACLs, firewalls) into one design question. It assumes the L2 material on VLANs, trunk ports and inter-VLAN routing.

Where to go next

References

  • NIST — Guide to Operational Technology (OT) Security (SP 800-82r3, September 2023) — https://doi.org/10.6028/NIST.SP.800-82r3 — the VLAN’s scope and the isolation devices (lines 11049–11050, 11059–11060, 5276–5277); the segmentation criteria (lines 3879–3880); the zone as a grouping precursor and the protected communications between zones (lines 3898–3902) and as a control boundary (line 9324); boundary devices do not mitigate lateral movement inside a zone (lines 4046–4049).
  • NIST — Guide to a Secure Enterprise Network Landscape (SP 800-215, November 2022) — https://doi.org/10.6028/NIST.SP.800-215 — segments from VLAN IDs, policies from addresses (lines 1193–1195); the perimeter model’s failure and lateral movement (lines 492–494); micro-segmentation’s purpose (lines 1130–1131).
  • NIST — Guidelines on Firewalls and Firewall Policy (SP 800-41 Rev. 1, September 2009) — no DOI is printed in the converted text used here, so the publication is cited by number — the firewall at a logical network boundary (lines 942–943); differing security requirements (lines 934–937); the limit of inspection (line 893); policy coordination for nested boundaries (lines 1082–1083).
  • NIST — Zero Trust Architecture (SP 800-207, August 2020) — https://doi.org/10.6028/NIST.SP.800-207 — the implicit trust zone (lines 483–484, 492–493); the enforcement point closer to the resource (lines 495–496).
Nodrius Field Kit — cover

Get the Nodrius Field Kit

Enter your email and we'll send you the complete Field Kit: eight technical resources in one download.

Your email address is sent to Nodrius and used to deliver the Field Kit to you by email. Marketing messages are separate and optional: the checkbox does not affect your download, and you are only added to our updates list if you tick it. Privacy Policy.

Privacy & cookies

This site does not use cookies, analytics or tracking, and it does not profile you. There is nothing technical to switch off, so Accept and Reject change nothing about how the site works: either choice lets you browse everything normally.

The only difference is what your browser remembers: two entries in local storage — your choice here, so this notice is not shown again, and a marker that the Field Kit offer has already been made — so the offer is presented once and not on every page. They contain no identifier, they are not shared with anyone, and the site sets no cookie for either.

What we do with your data
How your email address is used when you ask for a Resource, how long it is kept and which rights you have — in the Privacy Policy.
What the site stores in your browser
Two entries in local storage — your choice on this notice and whether the Field Kit offer has been shown — and no cookies, no session storage, no third-party content: in the Cookie Policy.