PrintNightmare and the Patch Tuesday rhythm

The PrintNightmare disclosures (CVE-2021-1675 originally, CVE-2021-34527 as the broader chain that emerged from the research-community analysis through July, and the various derivative issues that have continued through August into September) have produced more sustained operational work across the customer-portfolio Windows estate than any single 2021 vulnerability series so far (Microsoft Security Response Center on PrintNightmare, July through September advisories). The substantive operational lesson is about Windows Print Spooler specifically and about service-attack-surface management generally.

The Print Spooler service has, on the public-vulnerability record over many years, been a recurring source of critical-severity findings. The service runs by default on Windows endpoints, has historically had elevated privileges to support legacy printer-driver installation patterns, and has a substantial codebase that has not consistently received the security-engineering attention that more visibly user-facing components have. The PrintNightmare chain — pre-authentication remote code execution against Windows hosts running Print Spooler with default configuration — is one in a series of comparable findings against the same component. The structural answer is, on the operational evidence, to disable the Print Spooler service on hosts that do not require it, which is most servers and many endpoints. The cost of that disablement is low for hosts that do not need to print; the security benefit is substantial.

For the customer-portfolio response over the summer, the action has been the systematic disablement of Print Spooler on customer-organisation servers and on workstations that do not require local printing. The customer-organisation IT-operations conversations about this have been more substantive than I would have predicted — disabling a default Windows service that users are not actively asking to use is, organisationally, a less-friction action than the customer-organisation IT-operations function had treated it as for years. The PrintNightmare findings have produced the catalysing moment for the disablement to be operationalised at scale.

The wider strategic point is about service-attack-surface management as a discipline. Default Windows installations include many services that are running for legacy compatibility reasons, that have a non-trivial attack surface, and that are not actively used by most user populations. The disciplines for systematically reducing that attack surface — service-by-service review against the customer-organisation actual usage patterns, automated configuration management to enforce the reductions across the estate, and continuous validation that the reductions remain in place across endpoint refresh and OS update cycles — are operationally tractable but require sustained programme attention. The customer-organisation programmes have, on the PrintNightmare-driven momentum, been investing in this discipline more substantively through 2021 than in previous years.

For the EmilyAI detection content, the PrintNightmare-related TTPs have been incorporated continuously through the summer. The detection coverage on customer estates against the specific exploitation patterns is comprehensive. The post-disablement audit cycle has, on a small subset of customer hosts, found cases where the Print Spooler disablement was reverted by subsequent endpoint refresh or imaging cycles, and the configuration-management hardening to prevent that reversal has been part of the customer-organisation programme work.

I will write less on this specific chain because the operational story is, by mid-September, well-understood and the customer-portfolio work is substantively in motion. The structural lessons are documented and the follow-on work is routine. The chain is, however, a useful contemporaneous example of what 2021's continuous-vulnerability-disclosure rhythm looks like in operational practice — there is no week without substantive patching work, the customer-portfolio cadence is now operationally normalised against the rhythm, and the ability to maintain effective response without operational fatigue is the team-management discipline that has, over the past several years, become more central than the individual technical response.


Back to all writing