Compensating Controls When a Patch Cannot Be Deployed

A critical vulnerability can become a production incident the moment the change window closes. A core banking service, hospital system, industrial controller or mining platform may need to keep running while the vendor patch is still in testing, depends on another release or cannot tolerate a reboot.

The answer is not to leave the exposure unattended. Teams can apply compensating controls, meaning temporary measures that block, narrow or detect the likely exploit path until the permanent fix is deployed. Done well, this reduces business interruption, creates evidence for regulators and insurers, and gives engineering teams a controlled route to remediation rather than a vague exception.

Why Patches Sometimes Miss the Window

Patch deferral is usually an operating-model problem, not a lack of security intent. Legacy equipment may run unsupported firmware, vendor release trains may not match an organisation’s customisations, or application components may depend on a certification that takes months. Reboot requirements can also turn a small component update into a full service outage.

Those constraints are familiar across Australia. A FIFO-operated Pilbara site may have limited local technical staff and scarce satellite bandwidth, while a Sydney payments platform cannot be rebooted during peak settlement. Year-end change freezes and operational commitments narrow the window further. For a team in Melbourne or Brisbane, a patch scheduled overnight can land during the afternoon in Perth, when the asset is back in full use and business as usual has resumed.

A partner-managed system adds another approval layer, but it does not transfer accountability for the underlying risk. Current intelligence on exploited flaws can help teams separate a theoretical scanner result from an active threat. The CARM Threat Centre is a useful starting point for comparing campaigns and vulnerability disclosures with the organisation’s own asset inventory.

Frame the Deferral as Managed Risk

Before approving a deferral, state exactly what remains exposed, who could exploit it, which business services depend on the asset and how long the exception will last. Separate technical facts from commercial consequences: remote code execution on a domain controller is different from the same scanner result on an isolated test device. A concise threat scenario keeps the decision grounded.

Prioritise using exploitability, exposure, asset criticality and the strength of existing safeguards. Vendor advisories and exploit-prediction data can sharpen the ranking, but they are inputs rather than verdicts. The ASD’s Essential Eight places patching alongside application control, multifactor authentication and regular reviews, so a missing patch should be assessed as part of the wider control environment.

In Australia, the ownership and evidence need to fit the applicable obligations. A financial institution may need to document vulnerability and patch management under APRA CPS 234; an agency handling personal information must consider the OAIC’s Notifiable Data Breaches scheme and privacy obligations. An exception does not suspend those duties. The record should name the accountable executive or service owner, explain why the risk was accepted and set a review date.

Reduce the Exploitable Attack Surface

The strongest interim measure is often subtraction, not another security product. Remove unnecessary public exposure, disable the affected feature, close obsolete ports or restrict protocols. If a service does not need inbound internet access, enforce that requirement at the firewall, cloud security group and application routing layer rather than relying on a reminder for administrators.

Segmentation limits lateral movement from a compromised endpoint to a vulnerable server. Use micro-segmentation, host firewalls and identity-aware access controls so only required workloads and administrators can reach the asset. Pair network restrictions with least-privilege roles, just-in-time administration and phishing-resistant multifactor authentication for privileged access. Attackers often pivot through trusted pathways, so the route to critical systems matters as much as the original exploit.

At the application layer, a targeted web application firewall rule can block a known exploit string or suspicious request pattern. A virtual patching service can offer broader protection while the real update is tested. Neither replaces correcting the vulnerable code: false positives, incomplete signatures and changing exploit techniques will erode their value. Apply them as short-lived, observable controls with a named expiry date.

Select Interim Safeguards by Layer

Choose controls that address the credible exploitation path for the specific product and version. A layered set usually works better than a single expensive purchase, provided each layer has a clear purpose and an accountable owner. Confirm that it can be deployed without disrupting the service or creating an unmanaged back door through exceptions.

For network and identity layers, use measures that reduce reachability and privilege:

For application and endpoint layers, add controls that constrain code execution and improve detection:

Do not apply every control to every asset. That creates operational drag, alert fatigue and maintenance debt. Order the work by exploitability and blast radius: immediate exposure reduction first, detection and forensic visibility next, then durable architectural changes. Give each control a rollback plan and an owner responsible for reviewing its effectiveness.

Test, Monitor and Document

Treat each temporary control as a change that needs validation. Test a firewall rule in logging mode before enforcement where possible, confirm that legitimate traffic still works and review privileged access against actual service accounts. Run an exploit simulation only with suitable authorisation and a safe environment; the objective is to prove the mitigation blocks the route, not to breach production.

Monitor control effectiveness and signs of attack together. Track blocked requests, authentication attempts, process activity and outbound connections; record alerts tied to the affected asset, rather than leaving them in a generic dashboard. Set a review cadence according to risk, and define an escalation trigger such as an attempted exploit or evidence that a control is being bypassed.

Keep the exception record with the service’s risk documentation. Include the vulnerability identifier, asset owner, threat scenario, selected controls, validation evidence, residual risk, approval, review date and patch dependency. This package supports assurance reviews under frameworks such as the ASD’s Essential Eight and gives management a clear view of the exposure that remains.

Coordinate People, Vendors and Systems

A vulnerable asset rarely belongs to one team. The service owner knows the business impact, operations knows the change constraints, security owns detection and response, and procurement or the supplier may control the fix. Name a coordinator for each exception and give them one deadline, one risk statement and one escalation path. Otherwise, a patch request can wait in a queue while the interim control quietly ages.

Engage the vendor early for a supported mitigation, feature-disablement plan or an earlier release commitment. Ask for affected versions, recommended safeguards and the date the permanent fix will be validated. Independent penetration testers, managed detection and response providers and local managed service providers can add capacity, especially for regional organisations that cannot staff a 24/7 security desk internally.

The response plan must also cover a compromise before the fix arrives. Define who can isolate the asset, how forensic data will be preserved, and how affected credentials, customers or records will be handled. CARM Security, developed by Exclusive Networks, supports coordinated post-breach attack remediation and mitigation by bringing vendor capabilities into a common response process. That matters because prevention work should not delay containment if exploitation is already under way.

Move From Exception to Remediation

No interim control should become permanent by default. Compare the available options on deployment effort, attack-path coverage, operational impact and sustainability. A measure that needs manual rule changes every Friday may be acceptable for a short emergency, but it is a weak foundation for a long-lived asset. Record the funding, full patch plan, validation criteria and firm review date.

Interim control Typical rollout Attack-path reduction Important limitation
Firewall rule for a known exploit Hours Blocks a specific request pattern or destination Evolving signatures and false positives
Network segmentation or access-control change Days to weeks Limits reachability and lateral movement Dependencies and change coordination
Application allowlisting Days Restricts executable code on an endpoint Compatibility and policy tuning
Detection and response tuning Days Improves visibility of suspicious activity Does not prevent exploitation alone
Feature or protocol disablement Hours Removes an exploitable capability May affect required business functions

Set a maximum period for each exception according to risk, then review it against the original goal: deploying the permanent patch. If the vendor date slips, escalate the decision instead of extending the deadline automatically. Accepting a longer deferral may be reasonable when the attacker path is closed, but it should be an explicit, evidenced choice by the accountable executive or risk owner.

The practical takeaway is to treat every patch deferral as a temporary engineering task with a measured attack path, deployed safeguards, validation evidence and an end date. That approach keeps essential Australian services running while making the remaining exposure visible, reviewable and far harder for an attacker to exploit.