Q-Cyber Magazine
Zero trust for government networks: From perimeter habit to per-request decisions
NIST SP 800-207 and CISA's Zero Trust Maturity Model 2.0 give public authorities a shared language for moving from location-based trust to least-privilege, per-request access.

Zero trust is not a product category. It is a change in the question government networks ask before they grant access: not "where is this request from?" but "is this user, device and action entitled, right now?"
What the definition actually changes
Government networks were built around a familiar habit: treat the inside of a ministry or data centre as safer than the outside, and invest in the boundary. NIST SP 800-207 records why that habit has worn thin. Users are remote, devices are mixed, and a growing share of assets sit outside an enterprise-owned network. Network location is no longer a reliable proxy for trust.
The implication is precise. Protection should focus on resources — systems, services, workflows and data — rather than on network segments. Access decisions should be as granular as practicable, and they should be made with the working assumption that the network itself may already be compromised. CISA's public description of the model uses the same idea: minimise uncertainty when enforcing accurate, least-privilege, per-request decisions.
This is a cultural change as much as a technical one. A virtual private network that simply extends the old inside to a laptop at home is not, by itself, a zero trust architecture.
Exhibit 1
CISA Zero Trust Maturity Model 2.0: pillars and cross-cutting capabilities
Source: Qualitative planning matrix derived from CISA, Zero Trust Maturity Model Version 2.0, which specifies five pillars and three cross-cutting capabilities, each with traditional, initial, advanced and optimal examples. No maturity scores are assigned.
A government reading of the maturity model
CISA presents the model as one of several road maps, not as a certification. Version 2.0 aligns to the US Office of Management and Budget memorandum M-22-09. Indian departments are not bound by that memorandum, but they can still use the four-stage language — traditional, initial, advanced, optimal — as an honest self-assessment. Most government estates will find themselves uneven: stronger on network perimeter than on data classification, or stronger on identity for staff than on service-to-service authentication.
The right next step is rarely a single platform purchase. It is to pick one or two pillars where implicit trust is still the default — often shared file rooms, flat administrative networks, or standing privileged accounts — and to replace that default with a per-request rule.
Implications
- Describe the resource, not the subnet. Every significant government system should have an owner, a classification and an access policy that does not collapse to "inside the WAN."
- Make identity and device posture visible together. A valid password on an unmanaged or unhealthy device is not a complete decision.
- Segment with purpose. Network architecture remains relevant, but as a way to enforce policy, not as a synonym for trust.
- Govern exceptions. Zero trust fails quietly when emergency access paths are undocumented and never expire.
What this means for government
A department does not reach zero trust by announcement. It reaches it by retiring implicit trust, one resource at a time, against a published maturity model. NIST SP 800-207 and CISA's five-pillar map are sufficient to start that work this quarter.


