Control families: Where hosting ends and SaaS begins
A practical way to understand responsibilities is to walk through major NIST SP 800‑53 families:
Access control (AC)
Platforms support IAM, but the vendor must design roles, enforce least privilege, implement MFA for privileged functions, manage session controls, and run periodic access reviews. Joiner, Mover, Leaver processes must be documented and automated where possible.
Audit and accountability (AU)
Infrastructure logging is useful, yet agencies need application events: admin actions, configuration changes, data exports, permissions changes, and failed access attempts. Vendors must define log retention, tamper resistance, correlation with platform logs, and alerting on high-risk events.
Configuration management (CM)
Baseline images and hardened OS builds may be inherited. The vendor owns application configuration baselines, change control, peer review, segregation of duties, and drift detection. Evidence includes change tickets, approvals, and rollback procedures.
Identification and authentication (IA)
The platform enables SSO and MFA, but the vendor must implement user lifecycle management, password/secret policies for service accounts, API authentication, and periodic credential hygiene checks.
Incident response (IR)
Cloud providers respond to infrastructure incidents. Vendors must maintain IR playbooks, conduct tabletop exercises, define customer notification thresholds/timelines, and perform post-incident reviews with documented corrective actions.
System and communications protection (SC) & media protection (MP)
Vendors decide and document cipher suites, TLS enforcement, certificate renewal, data flow protections, API security design, and token management. If customer-managed keys (CMK) are required, vendors must support and evidence key rotation and access controls.
Contingency planning (CP)
Backups can leverage platform services, but vendors must define RPO/RTO, test restoration, rehearse failover, document dependency maps, and maintain runbooks for disasters.
Personnel security (PS) & awareness and training (AT)
Vendors need background checks where applicable, role-based security training, phishing simulations, and policy acknowledgements for staff with system access.
These control families illustrate why a secure platform is necessary but is insufficient. Agencies must see design, operation, and evidence at the application tier.
Examples of non‑inherited controls
Common application owned‑ controls include:
- Incident response planning with roles, escalation paths, and customer communications.
- Authenticated vulnerability scanning of the application and rapid patch management.
- End-to-end encryption with documented key ownership and rotation.
- Role-based access control and access recertification for admins and sensitive data users.
- Continuous monitoring with dashboards, alerts, and monthly reporting.
- Configuration management with baselines, change tickets, and approvals.
- Backup and recovery drills with restoration evidence and failover rehearsal.
- Workforce training focused on secure operations and compliance responsibilities.
Agencies should request artifacts that prove these controls exist and operate, not just policy statements.
Case study 1: The logging gap
A finance agency adopts a SaaS analytics application hosted on an authorized platform. Infrastructure logs are collected, but the app only records authentication events—no admin actions, no configuration changes, no data export activity. During an audit, reviewers ask for evidence that access to sensitive dashboards is monitored and anomalous usage is detected. The vendor cannot provide correlated application tier logs. Result: a finding, a POA&M entry, and delayed rollout.
Resolution
The vendor must implement rich application logging (admin actions, permission changes, data exports), route logs to the agency SIEM, and set retention and alert rules. However, these changes alone are not sufficient for compliance. The fixes must be independently assessed by a 3PAO, documented in the System Security Plan (SSP), and included in continuous monitoring reports. The agency should review evidence of these controls and require ongoing verification, not just rely on vendor self-attestation.
Case study 2: Encryption ownership
A public health agency requires end-to-end encryption of patient identifiers with agency-controlled keys. The vendor relies on platform-managed keys and default rotation but lacks CMK support and clear certificate lifecycle procedures. A certificate expires unexpectedly, causing an outage and data availability risk.
Resolution
The vendor must enable CMK, document rotation schedules, automate certificate renewal, and create incident runbooks for key compromise scenarios. However, as with all compliance-related changes, these must be independently assessed by a 3PAO, documented in the SSP, and included in continuous monitoring. The agency should require evidence of these controls and ongoing verification, not just trust vendor claims.
Procurement workflow: Verifying compliance step‑by‑step
A structured workflow prevents surprises and accelerates time-to-value:
- Pre‑solicitation: Confirm vendor listing on FedRAMP Marketplace or GovRAMP directory. Identify the targeted impact level and any existing ATOs.
- RFP/RFI requirements: Require current authorization or a sponsored path; request artifacts (SSP, policies, test plans); specify monthly scan delivery and POA&M reporting cadence.
- Evaluation: Review system boundary, data flows, and inherited vs. vendor controls. Ask who owns encryption keys, how logs are correlated, and how identity is governed.
- ATO endorsement: Coordinate with the authorizing official to document conditions, constraints, and integration requirements (e.g., SIEM ingestion, SSO, incident playbooks).
- Implementation & handoff: Capture configuration baselines, logging destinations, backup schedules, and incident escalation timelines. Establish change control hooks.
- Continuous Monitoring: Track monthly scans and remediation SLAs; review annual assessments; require access recertifications and training attestations at agreed intervals.
Agency checklist
Use this checklist during procurement and oversight:
Marketplace and authorization
- Vendor appears on the FedRAMP/GovRAMP directory.
- Impact level matches data sensitivity (via FIPS‑199).
- SaaS product ATO status is documented and current.
Documentation and evidence
- SSP complete and current; boundary and data flows clearly drawn.
- Policies for AC, AU, CM, IA, IR, CP, SC, AT, PS approved.
- 3PAO results provided; open findings tracked in POA&M.
Technical controls
- MFA enforced for admin access; least privilege documented.
- Application logging configured; retention and SIEM correlation in place.
- Encryption end-to-end; certificates, keys, and secrets managed securely.
- Configuration baselines defined; change control procedures operational.
Operations and monitoring
- Monthly vulnerability scans shared; remediation within SLAs.
- Incident response plan tested; customer notification timelines documented.
- Backup and restoration tested; RPO/RTO defined; failover rehearsed.
- Training and awareness conducted; access recertification performed quarterly.