Skip to content

SOC 2 Trust Service Criteria - Cloud Control Mapping

Scope

Covers SOC 2 Trust Service Criteria (Security, Availability, Processing Integrity, Confidentiality, Privacy) mapped to cloud services, including audit preparation, evidence collection, and common audit findings. Does not cover SOX IT General Controls (see compliance/sox.md) or general governance patterns (see general/governance.md).

Overview

SOC 2 (System and Organization Controls 2) is an auditing framework developed by the AICPA (American Institute of Certified Public Accountants). It evaluates an organization's information systems based on five Trust Service Criteria (TSC): Security, Availability, Processing Integrity, Confidentiality, and Privacy.

Report Types: - Type I: Evaluates the design of controls at a point in time. - Type II: Evaluates both the design and operating effectiveness of controls over a period (typically 6-12 months).

Applicability: SOC 2 is not a regulatory requirement but is often demanded by enterprise customers as assurance of service provider security. It is especially relevant for SaaS, PaaS, and managed service providers.

Security (Common Criteria) is required. The other four criteria (Availability, Processing Integrity, Confidentiality, Privacy) are optional and included based on the nature of the service.


Security (Common Criteria) - Required

CC1: Control Environment

Purpose: The foundation of all other controls. Demonstrates that the organization is committed to integrity, ethical values, and effective governance.

Cloud Service Mapping

Control AWS Azure GCP
Governance structure AWS Organizations, Control Tower Management Groups, Azure Lighthouse Resource Manager, Organization Policy
Policy enforcement SCPs, Config Rules Azure Policy, Blueprints Organization Policy, Assured Workloads
Compliance evidence Audit Manager, AWS Artifact Compliance Manager, Service Trust Portal Compliance Reports, Assured Workloads

Architect Checklist

  • [Recommended] Organizational structure and reporting lines for security are documented
  • [Recommended] Board/management oversight of security program is evidenced
  • [Recommended] Code of conduct and ethics policy covers technology and data handling
  • [Recommended] Cloud governance framework (Organizations, Management Groups, Resource Manager hierarchy) is implemented
  • [Critical] Compliance monitoring dashboards are configured (Audit Manager, Compliance Manager)

CC2: Communication and Information

Purpose: The organization obtains, generates, and uses relevant quality information to support internal controls. Internal and external communication supports controls.

Cloud Service Mapping

Control AWS Azure GCP
Security notifications SNS, Security Hub findings Service Health, Defender alerts Cloud Monitoring alerts, Advisory Notifications
Documentation Internal wikis, runbooks Internal wikis, runbooks Internal wikis, runbooks
Status communication Health Dashboard Azure Status, Service Health Google Cloud Status Dashboard

Architect Checklist

  • [Recommended] Security policies are documented, approved, and distributed to relevant personnel
  • [Recommended] Security alert notifications are configured and routed to appropriate teams
  • [Recommended] External communication channels exist for security disclosures and incident notifications
  • [Recommended] System changes are communicated to affected stakeholders
  • [Recommended] Cloud provider security bulletins and advisories are monitored

CC3: Risk Assessment

Purpose: The organization identifies and assesses risks to the achievement of its objectives, including fraud risk.

Cloud Service Mapping

Control AWS Azure GCP
Risk identification Security Hub, Inspector, GuardDuty Defender for Cloud, Secure Score Security Command Center, Security Health Analytics
Vulnerability assessment Inspector, ECR scanning Defender Vulnerability Management Artifact Analysis, Web Security Scanner
Threat intelligence GuardDuty threat feeds Threat Intelligence (Sentinel) Chronicle Threat Intelligence
Configuration risk AWS Config, Trusted Advisor Azure Advisor, Policy compliance Security Health Analytics, IAM Recommender

Architect Checklist

  • [Recommended] Formal risk assessment process is established and conducted at least annually
  • [Recommended] Risk register includes cloud-specific risks (misconfiguration, data exposure, account compromise)
  • [Critical] Vulnerability scanning is automated and continuous
  • [Recommended] Threat detection services are enabled across all accounts/subscriptions/projects
  • [Recommended] Risk tolerance levels are defined and drive control implementation decisions
  • [Recommended] Third-party and supply chain risks are assessed (including cloud provider risk)

CC4: Monitoring Activities

Purpose: The organization monitors internal controls and evaluates their effectiveness on an ongoing basis.

Cloud Service Mapping

Control AWS Azure GCP
Continuous monitoring Security Hub, Config, CloudWatch Defender for Cloud, Monitor, Policy Security Command Center, Cloud Monitoring
Compliance monitoring Audit Manager Compliance Manager Assured Workloads
Deficiency tracking Security Hub findings, Jira/ticketing Defender recommendations, ticketing SCC findings, ticketing
Control effectiveness Config conformance packs Policy compliance state Organization Policy audit

Architect Checklist

  • [Recommended] Continuous monitoring is implemented across all cloud environments
  • [Recommended] Security Hub / Defender for Cloud / Security Command Center aggregates findings
  • [Recommended] Control deficiencies are tracked in a ticketing system with SLAs for remediation
  • [Critical] Compliance drift is detected and alerted upon
  • [Recommended] Management reviews monitoring results at defined intervals
  • [Recommended] Independent evaluations (internal audit, external assessments) supplement ongoing monitoring

CC5: Control Activities

Purpose: The organization designs and implements control activities to mitigate risks to acceptable levels.

Cloud Service Mapping

Control AWS Azure GCP
Infrastructure as code CloudFormation, CDK, Terraform ARM/Bicep, Terraform Deployment Manager, Terraform
CI/CD controls CodePipeline, CodeBuild Azure DevOps, GitHub Actions Cloud Build, Cloud Deploy
Change management CloudFormation change sets, CodePipeline approvals Azure DevOps approvals, ARM what-if Cloud Build approvals
Segregation of duties Separate accounts (dev/staging/prod), IAM Separate subscriptions, RBAC Separate projects, IAM

Architect Checklist

  • [Recommended] Infrastructure is deployed via code (IaC) with version control and peer review
  • [Recommended] CI/CD pipelines enforce quality gates (testing, security scanning, approval)
  • [Recommended] Change management process requires approval before production deployment
  • [Critical] Segregation of duties is enforced (developers cannot deploy to production unilaterally)
  • [Recommended] Environment separation (dev, staging, production) prevents unauthorized access to production data
  • [Recommended] Automated controls are preferred over manual controls for consistency
  • [Recommended] Technology controls are mapped to specific risks they mitigate

CC6: Logical and Physical Access Controls

Purpose: The organization restricts logical and physical access to authorized users and protects system boundaries.

Cloud Service Mapping

Control AWS Azure GCP
Identity management IAM, IAM Identity Center, Cognito Entra ID, Entra External ID (formerly Azure AD B2C) Cloud Identity, Identity Platform
MFA IAM MFA, Identity Center MFA Entra MFA, Conditional Access 2-Step Verification, Titan Keys
SSO / federation IAM Identity Center (SAML, OIDC) Entra ID (SAML, OIDC, WS-Fed) Cloud Identity (SAML, OIDC)
Network access controls Security Groups, NACLs, VPC NSGs, Azure Firewall, VNet Firewall Rules, Cloud Armor
Encryption at rest KMS, CloudHSM Key Vault, Managed HSM Cloud KMS, Cloud HSM
Encryption in transit ACM, ALB/CloudFront TLS Key Vault certs, App Gateway TLS Certificate Manager, LB TLS
Physical security AWS data center certifications Azure data center certifications Google data center certifications
Secrets management Secrets Manager, Parameter Store Key Vault Secret Manager

Architect Checklist

  • [Recommended] Identity provider is centralized (SSO/federation for all cloud accounts)
  • [Critical] MFA is enforced for all human users (phishing-resistant methods preferred)
  • [Critical] Least privilege is enforced -- permissions are scoped to specific resources and actions
  • [Critical] Service accounts use short-lived credentials (STS, managed identities, workload identity)
  • [Critical] Network access follows defense-in-depth (security groups + NACLs/firewall rules)
  • [Critical] All data is encrypted at rest using customer-managed keys for sensitive workloads
  • [Recommended] All data in transit uses TLS 1.2+
  • [Recommended] Secrets are stored in dedicated services and rotated automatically
  • [Recommended] Access is reviewed and recertified at least quarterly
  • [Recommended] Physical access is managed by cloud provider (verify SOC 2 report from provider)
  • [Recommended] Deprovisioning removes access within defined SLA (same day recommended)
  • [Critical] Root/global admin accounts are secured with hardware MFA and monitored for usage

CC7: System Operations

Purpose: The organization detects and responds to security incidents and other anomalies.

Cloud Service Mapping

Control AWS Azure GCP
Threat detection GuardDuty Defender for Cloud, Sentinel Security Command Center, Chronicle
Vulnerability management Inspector, ECR scanning Defender Vulnerability Management Artifact Analysis
Incident response Security Hub, Detective, EventBridge Sentinel, Defender for Cloud Chronicle, Security Command Center
Monitoring and alerting CloudWatch, CloudTrail Monitor, Activity Log Cloud Monitoring, Cloud Logging
Patch management Systems Manager Patch Manager Azure Update Manager OS Patch Management
Anti-malware GuardDuty Malware Protection Defender for Endpoint Security Command Center

Architect Checklist

  • [Recommended] Threat detection is enabled in all environments (GuardDuty, Defender, SCC)
  • [Critical] Vulnerability scanning covers infrastructure, containers, and application dependencies
  • [Recommended] Patch management process ensures critical patches are applied within defined SLAs
  • [Critical] Incident response plan is documented, includes escalation procedures, and is tested at least annually
  • [Recommended] Security events are correlated in a SIEM for real-time analysis
  • [Recommended] Alerting thresholds are tuned to minimize false positives while catching true threats
  • [Recommended] On-call rotation ensures 24/7 coverage for security incidents
  • [Recommended] Post-incident reviews are conducted and lessons learned are implemented

CC8: Change Management

Purpose: The organization manages changes to infrastructure and software in a controlled manner.

Cloud Service Mapping

Control AWS Azure GCP
IaC and version control CloudFormation, CDK, CodeCommit ARM/Bicep, Azure Repos Deployment Manager, Cloud Source Repositories
Deployment pipelines CodePipeline, CodeDeploy Azure DevOps, GitHub Actions Cloud Build, Cloud Deploy
Change tracking CloudTrail, Config Activity Log, Change Analysis Cloud Audit Logs, Cloud Asset Inventory
Rollback CloudFormation rollback, CodeDeploy rollback ARM rollback, App Service slots Cloud Deploy rollback
Testing CodeBuild, Device Farm Azure Test Plans, Azure DevOps Cloud Build

Architect Checklist

  • [Recommended] All changes go through a formal change management process (request, review, approve, deploy, verify)
  • [Recommended] Changes are tracked in version control with meaningful commit messages
  • [Recommended] Infrastructure changes are deployed via IaC -- no manual console changes in production
  • [Recommended] Peer review (pull request) is required for all code and infrastructure changes
  • [Recommended] Automated testing validates changes before production deployment
  • [Recommended] Rollback procedures are documented and tested
  • [Recommended] Emergency change procedures exist with post-facto review and approval
  • [Critical] CloudTrail / Activity Log / Audit Logs capture all infrastructure changes
  • [Recommended] Unauthorized changes are detected and alerted via Config / Policy / Asset Inventory

CC9: Risk Mitigation

Purpose: The organization identifies, selects, and develops risk mitigation activities, including vendor management.

Cloud Service Mapping

Control AWS Azure GCP
Vendor risk assessment AWS Artifact (SOC reports, certifications) Service Trust Portal Compliance Reports Manager
Business continuity Multi-AZ, Multi-Region, AWS Backup Availability Zones, Azure Site Recovery Multi-region, Cloud Storage dual-region
Insurance / transfer Cyber insurance (organizational) Cyber insurance (organizational) Cyber insurance (organizational)

Architect Checklist

  • [Recommended] Risk mitigation strategies are defined for each identified risk (accept, mitigate, transfer, avoid)
  • [Recommended] Cloud provider's SOC 2 Type II report is reviewed at least annually
  • [Recommended] Complementary User Entity Controls (CUECs) from provider SOC reports are implemented
  • [Recommended] Vendor risk assessments are conducted for all critical third-party services
  • [Critical] Business continuity and disaster recovery plans are maintained and tested
  • [Recommended] Cyber insurance coverage aligns with identified residual risks
  • [Recommended] Risk acceptance decisions are documented and approved by management

Availability (A1)

A1.1, A1.2, A1.3: Availability Commitments

Purpose: The system is available for operation and use as committed or agreed (SLAs).

Cloud Service Mapping

Control AWS Azure GCP
High availability Multi-AZ deployments, Auto Scaling, ELB Availability Zones, Scale Sets, Load Balancer Regional MIGs, Cloud Load Balancing
Disaster recovery Multi-Region, Elastic Disaster Recovery, AWS Backup Azure Site Recovery, Geo-Redundant Storage Multi-region storage, Cross-region replication
Monitoring CloudWatch, Health Dashboard Monitor, Service Health Cloud Monitoring, Status Dashboard
Capacity planning Auto Scaling, Compute Optimizer Autoscale, Advisor Autoscaler, Recommender
DDoS protection Shield, Shield Advanced DDoS Protection Cloud Armor
CDN / edge CloudFront Front Door, CDN Cloud CDN
DNS failover Route 53 health checks, failover routing Traffic Manager, Front Door Cloud DNS, Traffic Director

Architect Checklist

  • [Recommended] SLAs are defined and documented for all customer-facing services
  • [Recommended] Architecture supports the committed SLA (Multi-AZ minimum for 99.9%+)
  • [Recommended] Auto-scaling is configured with appropriate min/max/desired settings
  • [Critical] Health checks and automated failover are configured
  • [Recommended] DDoS protection is enabled (Shield/DDoS Protection/Cloud Armor)
  • [Critical] Disaster recovery plan defines RPO and RTO and is tested at least annually
  • [Critical] Backup and restore procedures are documented and tested
  • [Recommended] Capacity planning is reviewed quarterly to prevent resource exhaustion
  • [Recommended] Incident communication procedures include customer notification for availability events
  • [Recommended] Dependencies are identified and monitored (third-party services, APIs)
  • [Critical] Runbooks exist for common availability scenarios (failover, scaling, recovery)

Processing Integrity (PI1)

PI1.1 - PI1.5: Processing Integrity Commitments

Purpose: System processing is complete, valid, accurate, timely, and authorized.

Cloud Service Mapping

Control AWS Azure GCP
Input validation API Gateway validation, Lambda input validation API Management validation, Function validation Apigee validation, Cloud Functions validation
Data quality Glue Data Quality, DynamoDB conditions Data Factory data flows, SQL constraints Dataflow, BigQuery constraints
Processing monitoring CloudWatch metrics, Step Functions Monitor metrics, Logic Apps Cloud Monitoring, Workflows
Queue / messaging integrity SQS (exactly-once), SNS Service Bus (exactly-once), Event Grid Pub/Sub (exactly-once)
Transaction integrity RDS/Aurora transactions, DynamoDB transactions SQL Database transactions, Cosmos DB transactions Cloud SQL transactions, Spanner
Error handling Dead letter queues (SQS, SNS), CloudWatch Alarms Dead letter queues (Service Bus), Monitor Alerts Dead letter topics (Pub/Sub), Monitoring Alerts

Architect Checklist

  • [Recommended] Input validation is enforced at all system boundaries (API gateway, application layer)
  • [Recommended] Data processing outputs are verified against expected results (reconciliation)
  • [Recommended] Error handling captures and reports processing failures without data loss
  • [Recommended] Dead letter queues capture failed messages for investigation and reprocessing
  • [Recommended] Processing completeness is monitored (record counts, checksums, reconciliation reports)
  • [Recommended] Transaction integrity is enforced for multi-step operations (ACID where applicable)
  • [Recommended] Processing SLAs (timeliness) are defined and monitored
  • [Recommended] Batch processing includes validation checkpoints and rollback capability
  • [Recommended] Data lineage is tracked for critical processing pipelines
  • [Recommended] Customers are notified of processing errors that affect their data

Confidentiality (C1)

C1.1, C1.2: Confidentiality Commitments

Purpose: Confidential information is protected throughout its lifecycle as committed or agreed.

Cloud Service Mapping

Control AWS Azure GCP
Data classification Macie, resource tagging Purview Information Protection, resource tagging Sensitive Data Protection (DLP), resource labels
Encryption at rest KMS (CMK), CloudHSM, S3 SSE Key Vault, Managed HSM, Storage Encryption Cloud KMS, Cloud HSM, CMEK
Encryption in transit ACM, TLS enforcement Key Vault certificates, TLS enforcement Certificate Manager, TLS enforcement
Data loss prevention Macie (S3), CloudWatch + custom Purview DLP, Defender for Cloud Apps Sensitive Data Protection API
Access controls IAM, S3 policies, Lake Formation RBAC, Storage ACLs, Purview IAM, BigQuery ACLs, VPC Service Controls
Data masking DynamoDB client-side, custom Dynamic Data Masking (SQL) BigQuery column-level security, Data Catalog
Secure deletion S3 lifecycle, KMS key deletion Blob lifecycle, Key Vault soft delete + purge Cloud Storage lifecycle, KMS key destruction
Network isolation VPC, PrivateLink, VPC Endpoints VNet, Private Link, Service Endpoints VPC, Private Service Connect, VPC Service Controls

Architect Checklist

  • [Recommended] Data classification scheme is defined (public, internal, confidential, restricted)
  • [Recommended] All resources are tagged/labeled with their data classification
  • [Recommended] Automated data discovery scans identify unclassified confidential data
  • [Critical] Encryption at rest uses customer-managed keys for confidential data
  • [Critical] Encryption in transit uses TLS 1.2+ for all confidential data flows
  • [Recommended] DLP controls prevent unauthorized exfiltration of confidential data
  • [Recommended] VPC Service Controls / PrivateLink / Private Link restricts data access to authorized networks
  • [Recommended] Data masking is applied when full access to confidential data is not required
  • [Critical] Data retention and deletion policies are enforced automatically via lifecycle rules
  • [Critical] Confidential data is purged when no longer needed (crypto-shredding via key deletion is acceptable)
  • [Recommended] Confidentiality obligations are defined in customer contracts and enforced technically

Privacy (P1 - P8)

P1: Notice and Communication of Objectives

Purpose: The entity provides notice about its privacy practices.

Architect Checklist

  • [Recommended] Privacy notice is accessible to data subjects before or at the time of data collection
  • [Critical] Notice includes types of data collected, purposes, retention, sharing, and individual rights

Purpose: The entity communicates choices available and obtains consent.

Cloud Service Mapping

Control AWS Azure GCP
Consent management Cognito (custom attributes), custom Entra External ID (custom flows), custom Identity Platform, custom
Preference storage DynamoDB, RDS Cosmos DB, SQL Database Firestore, Cloud SQL

Architect Checklist

  • [Recommended] Consent mechanisms are implemented for data collection (opt-in where required)
  • [Recommended] Consent preferences are stored and enforced programmatically
  • [Recommended] Consent withdrawal mechanisms are functional and trigger data handling changes

P3: Collection

Purpose: Personal information is collected only for identified purposes.

Architect Checklist

  • [Recommended] Data collection is limited to what is necessary for stated purposes (data minimization)
  • [Recommended] Collection methods are documented and consistent with privacy notice

P4: Use, Retention, and Disposal

Purpose: Personal information is used, retained, and disposed of in accordance with objectives.

Cloud Service Mapping

Control AWS Azure GCP
Retention policies S3 Lifecycle, DynamoDB TTL, RDS snapshot retention Blob lifecycle, Cosmos DB TTL Cloud Storage lifecycle, BigQuery partition expiration
Automated deletion S3 Lifecycle rules, Lambda-based cleanup Azure Functions + lifecycle Cloud Functions + lifecycle, Cloud Scheduler
Data inventory Macie, Config Purview Data Map Data Catalog, Cloud Asset Inventory

Architect Checklist

  • [Critical] Data retention periods are defined for each category of personal information
  • [Critical] Automated deletion enforces retention policies (lifecycle rules, TTL)
  • [Recommended] Personal information is not used for purposes beyond those stated in the privacy notice
  • [Recommended] Data inventory tracks where personal information is stored across all systems

P5: Access

Purpose: Data subjects can access their personal information for review and update.

Architect Checklist

  • [Recommended] Self-service mechanisms allow data subjects to view their personal information
  • [Recommended] Data subject access requests (DSARs) can be fulfilled within required timeframes
  • [Critical] Authentication verifies data subject identity before granting access to their data

P6: Disclosure and Notification

Purpose: Personal information is disclosed to third parties only as authorized, with notification.

Architect Checklist

  • [Recommended] Third-party data sharing is documented and aligned with privacy notice
  • [Recommended] Data processing agreements are in place with third parties receiving personal information
  • [Critical] Breach notification procedures comply with applicable privacy laws (GDPR 72 hours, CCPA, etc.)

P7: Quality

Purpose: Personal information is accurate and complete for its intended purpose.

Architect Checklist

  • [Recommended] Data validation ensures accuracy at point of collection
  • [Recommended] Mechanisms exist for data subjects to correct inaccurate information
  • [Recommended] Data quality checks are performed periodically

P8: Monitoring and Enforcement

Purpose: The entity monitors compliance with its privacy commitments and has procedures for inquiries and disputes.

Cloud Service Mapping

Control AWS Azure GCP
Privacy monitoring Macie, CloudWatch Purview, Compliance Manager Sensitive Data Protection, SCC
Access logging CloudTrail, S3 access logs Activity Log, storage analytics Cloud Audit Logs

Architect Checklist

  • [Critical] Privacy compliance is monitored continuously (automated scanning for PII exposure)
  • [Recommended] Access to personal information is logged and auditable
  • [Recommended] Privacy impact assessments (PIAs) are conducted for new systems processing personal information
  • [Recommended] Complaint and dispute resolution procedures are documented and accessible
  • [Recommended] Privacy program effectiveness is reviewed at least annually

SOC 2 Audit Preparation Checklist

Evidence Collection

  • [Critical] Cloud provider's SOC 2 Type II report is obtained (available via AWS Artifact, Azure Service Trust Portal, GCP Compliance Reports)
  • [Recommended] CUECs (Complementary User Entity Controls) from provider report are mapped to customer controls
  • [Recommended] All controls are documented with evidence of design and operating effectiveness
  • [Recommended] Policy documents are current (reviewed and approved within the audit period)
  • [Recommended] Access review evidence is collected (quarterly reviews with approvals)
  • [Recommended] Change management records demonstrate consistent process adherence
  • [Critical] Incident response testing evidence is available (tabletop exercises, simulations)
  • [Critical] Vulnerability scan and penetration test reports cover the audit period
  • [Recommended] Business continuity / DR test results are documented
  • [Critical] Training completion records are available for all in-scope personnel

Common Audit Findings to Prevent

  • [Recommended] No orphaned accounts (terminated employee access still active)
  • [Recommended] No overly permissive IAM policies (wildcard permissions)
  • [Critical] No unencrypted data at rest or in transit
  • [Critical] No missing MFA on privileged accounts
  • [Recommended] No gaps in logging (all regions, all accounts)
  • [Recommended] No undocumented exceptions to security policies
  • [Recommended] No missing evidence for control operation during the audit period

Common Decisions (ADR Triggers)

  • TSC scope selection — which Trust Service Criteria beyond Security (Availability, Processing Integrity, Confidentiality, Privacy) to include
  • Audit period and type — Type I vs Type II, audit period length (6 vs 12 months), auditor selection
  • Evidence collection architecture — automated evidence collection tooling, continuous compliance monitoring, screenshot vs API-based evidence
  • Access review process — quarterly review scope, approval workflow, access recertification tooling
  • Change management process — CAB structure, emergency change policy, evidence of approval for every change
  • Incident response program — incident classification, escalation procedures, tabletop exercise frequency
  • Vendor management — cloud provider SOC report review cadence, CUEC mapping, subservice organization monitoring
  • Logging and monitoring architecture — centralized logging, retention policy, alerting thresholds, anomaly detection
  • Business continuity and DR — BCP/DR testing frequency, test documentation, recovery time validation
  • AICPA SOC 2 — official AICPA SOC 2 reporting framework and Trust Service Criteria
  • AWS Artifact — download AWS compliance reports including SOC 2 Type II
  • Azure Service Trust Portal — access Azure compliance reports including SOC 2 Type II

See Also

  • general/security.md — General security controls and architecture patterns
  • general/governance.md — Cloud governance, tagging, and policy enforcement
  • compliance/sox.md — SOX IT General Controls (complementary for publicly traded companies)
  • compliance/iso-27001.md — ISO 27001 ISMS (often pursued alongside SOC 2)