I see this pattern regularly: an organisation migrates government workloads to the cloud, commissions a penetration test, receives a report with a handful of medium-severity findings, remediates them, and considers the security question answered. Six months later, they discover that their agency customer requires an IRAP assessment — and that the penetration test, while useful, does not satisfy the requirement.
A penetration test and an IRAP assessment answer fundamentally different questions. A penetration test asks: “Can an attacker exploit this system?” An IRAP assessment asks: “Is this system fit for purpose at the required classification level, as measured against the Australian Government Information Security Manual?”
Both are valuable. Neither replaces the other. But if your cloud environment processes, stores, or communicates Australian Government data at OFFICIAL:Sensitive or above, the IRAP assessment is what your customer will require — and a penetration test alone will not satisfy that requirement.
Understanding the Difference
The confusion between penetration testing and IRAP assessments is understandable. Both involve security professionals examining your system. Both produce reports with findings. But the scope, methodology, and purpose are entirely different.
A penetration test simulates an adversary. The tester attempts to exploit technical vulnerabilities in your network, applications, and infrastructure. The output is a list of vulnerabilities ranked by severity. This is enormously useful for identifying technical weaknesses — but it tells you nothing about whether your governance is adequate, whether your documentation is current, whether your personnel hold appropriate clearances, or whether your data sovereignty obligations are met.
An IRAP assessment evaluates your system holistically against the ISM. It covers technical controls (which may include penetration testing as one input), but also governance, documentation, personnel security, physical security, and operational processes. The output is a Security Assessment Report that provides a recommendation on the system's suitability for processing government data at the target classification.