Repository navigation
Replies: 10 comments 3 replies
|
I think allowing Class D offerings to run on Class C infrastructure is a positive step. It's a practical path, especially for providers already on commercial platforms or whose workloads need specialized infrastructure. But Class D assurance can't simply be inherited from a Class C platform, so the provider has to make up the gap at its own layer. FRC-CLD-CCH should say what that means. At minimum, federal data is encrypted with keys the provider holds outside the platform; plaintext never reaches the platform's managed services; control-plane actions like snapshots, key changes, network changes and log tampering are either blocked or detected independently; and logs and backups live outside the platform, with tested recovery from losing the platform account. The authorization package (the SSP for Rev5/CR26, the SDR for 20x) should also state when the platform underneath is Class C, along with the compensating controls and the remaining risk, so agencies can make an informed call. It would help to clarify whether FRC-CLD-CCH applies to Rev5/CR26 Class D offerings or only to 20x, and how the new FOCI and personnel KSIs apply to the platform's operators in this setup. With clear expectations for equivalent security, this could be a big benefit for CSPs and the agencies that rely on them. |
|
I am in agreement with almost all of these requirements. Some might believe these are difficult, but if you are using enough automation and AI assistance then it is actually simpler and more secure to achieve than legacy rev 5. Especially, if you consider the time when class D will be fully released there will be even better tools available to assist in achieving the certification. VER-EVA-IOC - I have seen this in practice and definitely useful exercise when discovering a past vulnerability existed in a codebase. Allowing flexibility on level of effort to search for IOCs is good too as long as CSPs document how they searched, as reviewing a years' worth of logs with a script is easier than hiring a forensics team to search for IOCs. All organizations have vulnerabilities at one point or another even if they don't admit it openly so to me this is a valuable check. FRC-FRA-FOD - The FOCI ones are useful and important going forward as it provides better visibility into the overall setup of an organization. I agree with the setup too where it puts the details in the hands of the agency due to sensitivity of some information. It is clear adversaries have been using FOCI as a way in for a long time as seen in DOW posture to this problem, so it makes sense FedRAMP aligns. FOCI documentation is a positive step in transparency and a powerful security signal when used properly. FRC-CLD-PCD - Agree on the progressive steps to achieve class D. Gating the process is useful as we have seen spun up fake startups especially with AI attempt to map out the class D ecosystem. Any company that jumps right into class D without the proper experience should require more review. This allows a vetting period and ensures companies are actually ready and saves everyone time and money. FRC-CLD-CCH - Agree with this change, though it opens up an inheritance gap if CSPs don't architect around it properly. If a Class D CSP uses standard Class C serverless or managed PaaS options, an underlying platform support engineer (who may not meet Class D suitability expectations) could potentially access sensitive data or source code. KSI-IAM-MPI and KSI-IAM-MST should block this threat since a CSP typically cannot verify 3rd-party cloud support engineers persistently, the CSP should be forced to use technical controls like Confidential VMs, CMEK, and gated support access. FRC-CLD-CTP - Agree with this requirement too since it creates a cohesive secure system when the data can be tracked through vetted providers. It is scoped well to actual risk of federal data which makes it achievable for CSPs. Overall, these requirements are clear and reasonable for class D. |
|
My two cents: KSI-SVC-IED Inspecting Encrypted Data - I believe the intent here is endpoint decryption, but this should be clarified so it is not confused with a requirement for the provider to intercept and decrypt data in transit. It might also need to consider what measures are needed to ensure this ability does not weaken end-to-end encryption. SDR-CSX-KMT Key Security Indicator Metrics - Does this mean calling the API every 4 hours, or calling it and forcing the check to rerun every 4 hours? For example, would I just call AWS Config to get the latest status of an S3 bucket’s public access block setting, or would I need to rerun that check and then get the status? FRC-CLD-PCD Progressive Certification to Class D - What does “without corrective action” mean for the six-month Class C period? Is this limited to a formal FedRAMP Corrective Action Plan, or would an agency POA&M or a provider’s routine vulnerability remediation also reset the clock? |
|
Feedback on FRC-CLD-PCD: give active Rev5 Class D offerings a direct path to 20x Class D Why:
Suggested change: |
|
@pete-gov Is there a Class G offering now? Or is this just AI trying to create new things for fun? FRC-CLD-CTP Certified Third-Party Services Providers seeking a FedRAMP 20x Class D Certification SHOULD limit the likely handling of federal customer data within the cloud service offering to third-party information resources that are included in a FedRAMP Class B, Class C, Class D, or Class G Certified offering. Note: This is a NEW rule intended to ensure that the full lifecycle of federal customer data can be assessed for risk by agencies. |
|
A question on SDR-CSX-KMT, because the answer decides how often providers have to run KSI checks at Class D. The rule asks for "metric data (including status of persistent validation) every 4 hours." Does that set a 4-hour floor on how often KSI validation has to run? Right now nothing else does. FRD-PER says persistent activities "may occur irregularly, and may include interruptions or waiting periods between cycles." FRC-CSX-VVK asks for at least 4 automated methods per KSI at Class D but no interval. The tightest interval in the rules is VDR-TFR-MVX at 3 days, and this RFC proposes the same 3 days for Class D. If 4 hours is meant as a validation floor, please put it where the obligation lives: in FRC-CSX-VVK's Class D variant, or as a Class D note on FRD-PER. If it just means recorded status gets sampled every 4 hours, please say that, so a 3-day validation cadence doesn't look like a gap. |
|
Another question, from Boundera, on how the new VER-EVA-IOC fits with the incident rules we already have. At Class D, VER-TFR-IRI and VER-TFR-NRI already say to treat some vulnerabilities as FedRAMP Reportable Incidents: internet-reachable and likely exploitable at N4 or N5, and likely exploitable but not internet-reachable at N5. Once that happens, the ongoing incident reports (IEC-CSO-OIR) already ask for indicators of compromise. So is VER-EVA-IOC mainly for the vulnerabilities below that line? Take an internet-reachable, likely exploitable vulnerability at N3. It's not an incident, but a provider might still decide to search the logs. If they search and find nothing, that result has nowhere to go today. If that's the intent, it would help to say so in the rule, and add an optional field to the vulnerability detail record: IOC search done or not, when, and what it found. Then it's assessable without a new report. That also covers @sca1235's point about documenting how the search was done. |
|
Hi FedRAMP Team, CMU-CSO-PQS: Microsoft recommends removing the second part of this requirement because transition timelines already differ by customer type. FedRAMP would also expect any cryptographic module that implements PQC to complete FIPS 140 validation, and Microsoft can't give a meaningful date for when that validation would be done. Thank you |
|
Thank you FedRAMP Team for publishing RFC-0036 and advancing the FedRAMP 20x framework toward a more automated, evidence-driven approach to cybersecurity assurance. I believe there are some areas that still warrant further consideration:
Machine-readable certification evidence is valuable only when its authenticity, integrity, and origin can be independently verified. RFC-0036 requires API-accessible certification packages but does not explicitly establish the minimum trust requirements for the evidence exchanged. Recommendation: Consider requiring cryptographically verifiable evidence integrity, trusted timestamps, source attribution, and tamper-evident audit records. API access should also support strong authentication, granular authorization, and appropriate protections for sensitive security information.
Historical KSI reporting provides useful visibility into security posture. However, an indicator may appear compliant even when its underlying telemetry is incomplete, stale, or unavailable. Recommendation: Require KSI reporting to distinguish between compliant, noncompliant, unknown, and unavailable states. Include metrics for telemetry coverage, evidence freshness, and validation failures. Missing evidence should not automatically be interpreted as a successful security validation.
The proposed three-day verification requirement is a useful baseline. However, validating known resources does not necessarily demonstrate that all relevant resources have been discovered, particularly in environments using ephemeral workloads, containers, and dynamically provisioned infrastructure. Recommendation: Require CSPs to demonstrate reconciliation between discovered assets, cloud inventory, and deployment records. Consider an additional coverage metric identifying unmanaged or unverified resources.
Automation can improve consistency and scalability, but the existence of automated security tools does not necessarily demonstrate that the tools are operating effectively. Recommendation: Include periodic effectiveness testing of critical security automation, such as simulated detection failures, alert-response validation, and evidence of successful control execution. This would help distinguish operational security effectiveness from tool deployment alone.
I appreciate the explicit recognition of AI agents within the monitoring suitability requirement. As autonomous systems increasingly interact with federal cloud infrastructure, traditional identity and access controls may not sufficiently address delegated privileges, persistent credentials, and autonomous privileged actions. Recommendation: Clarify expectations for non-human identity ownership, privilege boundaries, credential lifecycle management, and traceability of privileged actions to an accountable identity or authorized workload. These safeguards should apply proportionately to the risks associated with each identity.
Without a common data structure, CSP-specific implementations may increase the integration and interpretation burden on federal agencies, undermining some of the efficiency gains intended by FedRAMP 20x. Recommendation: Publish a common, versioned KSI reporting schema with standardized fields for indicator identity, validation status, timestamps, evidence provenance, exceptions, and coverage. Where practical, align the schema with existing federal automation standards such as OSCAL.
Testing changes in non-production environments is a sound security practice. However, differences in infrastructure configuration, identity policies, network architecture, and third-party dependencies can limit the reliability of pre-production testing. Recommendation: Require CSPs to document material differences between production and testing environments and establish risk-based production validation for security-sensitive changes.
Given that Class D is intended to provide the highest level of assurance under FedRAMP 20x, the Phase 4 pilot should evaluate not only whether CSPs can implement the proposed requirements but also whether those requirements produce measurable security improvements. Recommendation: Establish and publish pilot evaluation criteria covering evidence quality, security control effectiveness, automation reliability, detection and response performance, agency usability, and operational burden. Use the findings to distinguish requirements that demonstrably improve assurance from those that primarily increase reporting complexity. |
|
FedRAMP PMO, Thank you for the opportunity to provide feedback on the draft RFC-0036. I highly support the PMO's modernization goals, particularly the shift toward automated validation, machine-readable artifacts, and deterministic telemetry. To help ensure these guidelines align seamlessly with modern, cloud-native engineering and Zero Trust Architectures, I respectfully submit the following constructive feedback for your consideration. 1. VER-TFR-MRH: Historical Activity
2. FRC-CLD-PCD: Progressive Certification to Class D
3. Key Security Indicators (KSIs): Numeric Quota of Tracking Methods
4. KSI-IAM-MST: Monitoring Suitability
I appreciate the PMO's dedication to collaborative modernization and look forward to partnering with the Board to support secure, high-assurance federal cloud environments. |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
RFC-0036 FedRAMP 20x Class D Cloud-Hosted Pilot Requirements
Note: You can discuss this RFC without making formal comment in discussion #181 .
Summary & Motivation
The FedRAMP 20x Cloud-Hosted Class D Pilot will explore the impact and implementation of the highest assurance level of FedRAMP Certification for cloud-hosted services. This RFC proposes specific changes to a variety of rulesets and key security indicators to test during the Phase 4 pilot.
This RFC includes the following:
Additional rules and Key Security Indicators may be added during or after the Phase 4 pilot, and the existing rules and indicators are likely to change based on direct feedback during the pilot. FedRAMP understands that it would be convenient to know exactly what the rules are likely to be a year in advance, but that is not a realistic approach to developing a new process that is responsive to real world impacts.
The addition of Foreign Ownership, Control or Influence (FOCI) information:
FedRAMP seeks to drive a balance between the amount of assurance related to FOCI that agencies might expect and the varying amounts of assurance related to FOCI that different cloud service providers may invest in. The goal is to provide basis rules and indicators that a cloud service provider can follow to provide an appropriate level of assurance, based on their own business strategy, in a fashion that can be shared with all potential agency customers. FedRAMP FOCI practices do not replace agency-specific needs or determinations, but should allow cloud service providers that implement them to save time and effort during an agency review.
FedRAMP will make no determinations on FOCI itself, only confirm whether cloud service providers adequately demonstrate that they persistently review the necessary indicators to understand FOCI and their business.
Feedback Requested
This RFC is intended specifically for the Phase 4 Cloud-Hosted pilot. FedRAMP is looking for feedback on the clarity, applicability, reasonability, and achievability of the proposed rules and Key Security Indicators for companies who wish to meet the highest levels of assurance for federal customers.
This RFC does not address Self-Hosted or Hybrid-Hosted rules or requirements as those are not part of the Phase 4 pilot.
New Definitions
FRD-FOR: Foreign Ownership, Control, Or Influence (FOCI)
FRD-FPS: Foreign Person
Miscellaneous Requirements
Please note that this RFC does not reproduce existing Class D rules that will apply to FedRAMP 20x in the Consolidated Rules for 2026; only new or modified rules are added. Participants are encouraged to review all existing Class D rules, of which there are many, for context while reviewing these updated requirements for FedRAMP 20x.
This section contains individual rules that are new or modified from existing rulesets.
Vulnerability Detection and Response
VDR-TFR-MVX Persistent Machine Verification and Validation for 20x
Vulnerability Evaluation and Reporting
VER-EVA-IOC Indicators of Compromise
VER-TFR-MRH Historical Activity
Security Decision Record
SDR-CSX-KMT Key Security Indicator Metrics
Certification Data Sharing
CDS-CSX-API API Certification Data
Cryptographic Module Use
CMU-CSO-PQS Post-Quantum Cryptography Status
FedRAMP Certification for 20x Class D
(Note: This section includes proposed formal language for some rules that were mentioned in RFC-0033, and the proposed language was adjusted based on feedback from that RFC)
FRC-FRA-FOD FOCI Determinations are Agency Specific
FRC-CLD-PCD Progressive Certification to Class D
FRC-CLD-CCH Certified Cloud Hosting
FRC-CLD-CTP Certified Third-Party Services
Key Security Indicators
FedRAMP has many rules that include specific assurance expectations for 20x (and Rev5) Class D compared to other FedRAMP Certification Classes. Key Security Indicators go beyond assurance to demonstrate that the cloud service provider is maintaining and tracking critical security information over time. FedRAMP 20x Class D providers will need to use at least 4 different automated methods to track Key Security Indicator status and are encouraged to go well beyond this. FedRAMP 20x Class D is intended to be competitive, such that providers who invest in additional automation, greater assurance, and high speed response and communication will consistently outperform providers that only seek to meet the minimal bar - though efficient providers at the minimum bar may compete on price or specific use cases.
In addition to all existing Key Security Indicators, FedRAMP proposes adding the following new Key Security Indicators specifically for FedRAMP 20x Class D based on the additional security capabilities expected from a High baseline (note that many automation capabilities are already addressed):
KSI-CMT-TNP Testing in Non-Production Environments
KSI-IAM-MST Monitoring Suitability
KSI-IAM-MPI Managing Personnel Identities
KSI-RPL-MCA Maintaining Cryptographic Access
KSI-SVC-IED Inspecting Encrypted Data
KSI-SVC-UAS Using Automated Security
KSI-PIY-DFC Detecting FOCI Changes
KSI-PIY-RFP Reviewing Foreign Persons
KSI-PIY-RFB Reviewing Foreign Businesses
KSI-PIY-MFD Maintaining FOCI Determinations or Mitigations
All reactions