Skip to content

Network vs Access Segmentation Principles

Network segmentation and access segmentation are complementary but distinct control philosophies. Understanding the difference — and when to apply each — is fundamental to secure architecture.

Network segmentation

Network segmentation partitions the network into isolated segments using Layer 2 (VLANs) and Layer 3 (firewalls, ACLs, routing) controls. Traffic between segments must pass through an enforcement point.

Controls: VLANs, firewalls, routers, ACLs, network zones
Policy anchor: Source/destination IP, subnet, port, protocol
Enforced at: Network infrastructure layer
Granularity: Subnet or VLAN level

Strengths

  • Works for all devices regardless of OS or agent support
  • Mature tooling and well-understood operational model
  • Effective for coarse-grained zone isolation

Weaknesses

  • East-west traffic within a segment is uncontrolled
  • Flat subnets allow lateral movement once inside a segment
  • Policy is tied to IP addresses — difficult to maintain as workloads move

Access segmentation

Access segmentation controls which users and devices can access which resources, independent of network location. The policy plane is identity, not topology.

Controls: IAM, RBAC, ABAC, ZTNA gateways, identity-aware proxies
Policy anchor: User identity, device posture, application, time, location
Enforced at: Application / identity layer
Granularity: Per-user, per-application, per-request

Strengths

  • Policy travels with the user regardless of where they connect from
  • Least-privilege access is easier to express and enforce
  • Supports continuous authorisation (re-evaluate mid-session)

Weaknesses

  • Requires agent support or browser-based access for enforcement
  • Legacy applications may not integrate with identity-aware controls
  • More complex policy management (identity lifecycle, JIT provisioning)

Comparing the models

DimensionNetwork segmentationAccess segmentation
Policy basisIP/subnet/portIdentity/context
Lateral movement controlPartial (zone boundaries)Strong (per-application)
Legacy device supportFullLimited without agents
Remote access fitPoor (VPN extends segment)Excellent (ZTNA)
East-west visibilityRequires additional toolingApplication-level by design
Operational maturity requiredLow-mediumMedium-high

The layered approach

Neither model alone is sufficient for high-security environments. The recommended approach layers both:

  1. Network macro-segmentation — zone isolation at the perimeter and between major tiers
  2. Network micro-segmentation — workload-level east-west control within zones
  3. Access segmentation — identity and device-aware policy for user access to applications
  4. Data segmentation — classification-based data access controls regardless of network path

This layered model implements defence-in-depth across the control planes: network topology, workload identity, user identity, and data sensitivity.

Further reading

Released under the MIT Licence.