Why Defense and Critical Infrastructure Operators Can’t Rely on Cloud-Based Remote Access 

The case for sovereign remote access infrastructure in high-security environments.

Remote access is not optional in defense and critical infrastructure environments. Distributed systems require maintenance. OEM equipment requires support. Contractors require access. Incidents often require response at locations where no technician is present.

The operational requirement is clear. The problem is the architecture used to meet it.

Over the past decade, the remote access market has consolidated around SaaS delivery models: vendor-managed portals, cloud-hosted control planes, and relay infrastructure operated by third parties. For most enterprise IT environments, that model works. For defense contractors, classified program offices, critical infrastructure operators, and organizations operating under strict data sovereignty mandates, it introduces a structural problem that no configuration option resolves.

The question is control, and specifically, what happens to it when remote access depends on infrastructure the organization does not own.

The Air Gap Was Never the End State

Air-gapped networks were designed around a specific premise: eliminating external connectivity eliminates the attack surface. No internet access. No external service dependencies. No path in from outside the perimeter.

That premise holds for data at rest. It does not hold for organizations that still need to operate, maintain, and support the systems inside the perimeter. Defense contractors operating sensitive communications infrastructure, energy operators managing geographically distributed grid assets, and government agencies running classified program environments all have contractor relationships, OEM support agreements, and maintenance workflows that require some form of remote access to function at operational scale.

Air gapping the network was never intended to eliminate remote access entirely. It was intended to control it absolutely. The question most organizations have not fully answered is whether their remote access architecture actually delivers that control, or whether it reintroduces (through the back door) exactly the external dependencies the air gap was designed to eliminate.

Why Cloud Remote Access Is Architecturally Incompatible

SaaS remote access architecture requires extending trust beyond the organization’s infrastructure boundaries, and in defense and critical infrastructure contexts, that extension is not acceptable.

Session traffic traverses third-party infrastructure. In most cloud-based remote access architectures, connections are brokered through vendor-managed relay infrastructure. The session originates inside the customer’s environment, routes through the vendor’s cloud, and terminates at the remote endpoint. Even with end-to-end encryption, the traffic path passes through systems the customer does not control, cannot audit independently, and cannot guarantee will not be subpoenaed, breached, or subject to foreign jurisdiction.

Authentication depends on external services. Cloud remote access tools typically rely on vendor-hosted identity and authentication infrastructure. Where identity governance is a sovereignty requirement (where the question “who authenticated this session, and from where?” must be answerable entirely within the organization’s own systems), external authentication dependencies are a compliance gap, not a configuration detail.

Audit records may not be fully owned by the customer. Session logs, connection records, and audit data generated through vendor-hosted platforms are stored in environments outside the customer’s direct control. Under NIS2 Article 21 and equivalent frameworks, the organization must own the audit trail, not the supplier. Cloud storage of session data creates a compliance dependency that cannot be eliminated without changing the architecture.

The attack surface expands by design. Cloud-delivered remote access platforms are internet-facing infrastructure managing access to high-value targets. They represent a concentrated, publicly addressable entry point into the organizations they serve. The supply chain attack vectors that NIS2 Article 21 §2.d specifically targets (for example, compromised vendor credentials, uncontrolled access pathways, ungoverned third-party connectivity) are most easily exploited when the access infrastructure itself is hosted externally.

VPNs Don’t Resolve It Either

Organizations that recognize the limitations of SaaS remote access frequently fall back on VPN-based architectures. VPNs keep session traffic within the organization’s own network, but they do not solve the access governance problem. In environments where time-bound authorization, least privilege, and full auditability are mandatory, VPNs introduce structural vulnerabilities.

A VPN grants an authenticated user access to a network segment. It does not restrict what that user can do within that segment, which systems they can reach, or how long that access remains valid. A defense contractor granted VPN access to support a single OT system is, in practice, connected to a network that may contain far more than that system. Lateral movement, the capability that characterizes the most damaging supply chain compromises against critical infrastructure, is a direct consequence of that architecture.

NIS2 Article 21 §2.d is explicit on this point: supply chain access must be governed at the level of specific devices, specific time windows, and specific authorized tasks. A VPN provides none of that granularity. Under NIS2’s supply chain security framework, VPN-based access to critical systems constitutes uncontrolled access.

Why Remote Access Is NIS2’s Highest-Stakes Requirement?

Read the latest white paper and understand the risk of noncompliance

The Requirement: Sovereign Remote Access Infrastructure

Defense primes, critical infrastructure operators, and government agencies with strict data residency mandates have reached the same architectural conclusion: remote access must operate entirely within the organization’s infrastructure boundary.

In practice, this means a self-hosted control plane: a portal deployed inside the organization’s own network, private cloud, or sovereign cloud environment, with no external relay, no vendor-managed authentication dependencies, and no session data leaving the boundary the organization controls. Every component of the remote access infrastructure, such as identity, policy, logging, and session brokering, must be owned and operated by the customer.

This differs fundamentally from a cloud remote access tool configured for “private mode.” It is purpose-built infrastructure for environments where the question is not whether to connect, but whether the connection can be fully controlled, fully audited, and fully contained.

The operational requirements such an architecture must satisfy are specific:

  • Third-party vendor and contractor access that is time-bound, role-restricted, and scoped to specific devices, without VPN risk, without persistent credentials, and without any traffic leaving the sovereign boundary
  • Remote OT and SCADA maintenance that services industrial systems without bridging IT and OT environments or requiring inbound network exposure
  • Contractor provisioning with access that expires automatically at the end of the authorized window, eliminating the credential sprawl that accumulates when access management is handled manually
  • Incident response capability that can be activated quickly, scoped precisely, and shut down completely, without creating permanent access pathways

In defense and critical infrastructure environments, these are standard operating requirements. The architecture used to meet them determines whether the organization retains actual control or merely the appearance of it.

What This Looks Like with Netop

Netop’s Self-Hosted Portal and Sovereign Portal are designed for organizations that have determined cloud-based remote access is not viable for their environment.

The architecture is fully containerized and infrastructure-agnostic. It deploys on-premises, in a private cloud, or within a sovereign cloud environment, including AWS GovCloud or EU-region deployments that meet national data residency requirements. All session brokering, authentication, logging, and policy enforcement operate within the customer’s own infrastructure. No traffic routes through Netop’s systems. No session data is stored externally. The organization owns the keys, certificates, logs, and audit trail.

At the access control layer, the architecture enforces policy-governed, Zero Trust principles at a granularity SaaS and VPN-based tools cannot match:

  • Per-session role-based access control, scoped to specific devices or device groups, with permissions that cannot be extended without explicit re-authorization
  • Time-window enforcement, so contractor access expires automatically when the authorized period ends — not when someone remembers to revoke it
  • IP-fencing, restricting sessions to whitelisted geographic ranges and named contractor addresses
  • Four-eyes oversight for high-risk sessions, requiring a designated supervisor to accept the connection before it proceeds
  • Full session recording with tamper-proof storage, capturing all keyboard events, mouse movements, and screen activity in sequence, producing the forensic-grade audit trail that NIS2 Article 23 and equivalent frameworks require

For organizations with Intel vPro-enabled infrastructure, Netop’s integration with Intel AMT extends this architecture below the operating system level, enabling BIOS-level access and One-Click Recovery for devices that are unresponsive or offline — without physical intervention and without compromising the sovereign boundary.

Participate in an Exclusive Pilot Program: Self-Hosted Portal

Sovereign remote access infrastructure for defense and critical environments

The Design Difference

Most remote access vendors built their platforms for operational convenience: ease of connection, cross-platform compatibility, fast deployment. Cloud delivery was a natural extension of that design philosophy, and for the majority of their customers, it is the right fit.

Defense contractors, critical infrastructure operators, and government agencies operating under sovereignty mandates require a different answer. The relevant question is not which remote access tool is easiest to connect. It is the architecture that maintains controlled, auditable access over every session, every endpoint, and every third-party interaction, without creating external dependencies that undermine the security posture of the rest of the organization’s infrastructure.

Netop has operated in these environments for over 40 years. The deployment flexibility that characterizes the platform on-premises, cloud, hybrid, air-gapped, or fully self-hosted was built in from the start, because the environments Netop was designed for have always required it.

For organizations that have not yet addressed the architecture question: remote access infrastructure that depends on systems you do not control, data you cannot fully audit, or access pathways you cannot completely terminate puts control of that access in someone else’s hands. In defense and critical infrastructure environments, that is a risk the architecture itself should eliminate.

Sign up to stay up-to-date on Netop latest updates.

More Insights

Learn how Netop secures remote access across legacy and modern OT systems, bringing consistent access controls, vendor permissions and auditing...

Water and wastewater utilities operate some of the most distributed infrastructure in critical operations: treatment plants, pump stations, lift stations,...

Starting September 1, Netop Self-Hosted Portal gives organizations operating in critical, regulated, or air-gapped environments a way to maintain secure...