Skip to main content
In this lesson we cover managing IAM for the enterprise at scale: the problem space, organizational and governance considerations, policy and access management, monitoring, and automation patterns. The demo is useful for hands‑on practice, but this article establishes the conceptual and practical foundation for secure multi-account architectures.
A slide showing a multi-account AWS security diagram: an "Organization" feeding into "Centralized Security Controls" that branch to several "AWS Account" boxes. To the right is an AWS Certified DevOps Engineer Professional badge.
Why multi-account?
  • Without centralized management you cannot reliably enforce policies across an enterprise. You lose governance, telemetry, and observability.
  • Large organizations with many accounts frequently experience inconsistent security rules, poor auditability, and limited central visibility.
AWS Organizations provides a management (formerly “master”) account that groups accounts into Organizational Units (OUs). OUs let you organize accounts by business unit, environment (dev/qa/prod), project, compliance domain, or any other structure that fits your governance model.
A slide titled "Scenario – The Multi-account Challenge" showing a "Before" panel with several separate account ovals (Account 01–05) and an "After" panel with a centralised management box that connects via arrows to multiple account ovals (all labeled Account 01).
OUs, delegation, and centralized controls
  • OUs allow you to apply centralized controls across groups of accounts.
  • Use Service Control Policies (SCPs) to restrict what actions are possible organization‑wide; SCPs can only limit or deny—not grant—permissions.
  • Delegate administrative duties using delegated administrator roles in specific accounts while keeping central oversight.
A hierarchical diagram of AWS Organizations showing a Management Account at the top, two Organizational Units (A and B) each containing two member accounts (Account 01–04). Below are Service Control Policies (SCPs) and Delegated Administrators indicating centralized policy enforcement and delegated administration.
Integrations and telemetry AWS Organizations integrates with many services to help apply controls and centralize telemetry:
  • AWS Control Tower (landing zone and account provisioning)
  • IAM Access Analyzer (cross-account/public access detection)
  • AWS Config (resource inventory & rules)
  • AWS CloudTrail (API activity logs)
  • AWS Service Catalog (standardized products)
A slide titled "AWS Organizations – Multi-Account Management Framework" showing AWS services integrated with Organizations. It displays five pink icons labeled AWS Control Tower, IAM Access Analyzer, AWS Config, AWS CloudTrail, and Service Catalog.
Designing your OU structure Design should reflect business drivers: environment separation, business unit boundaries, compliance domains (PCI, HIPAA), function, or a hybrid. Keep inheritance and operational scale in mind—Organizations supports up to five levels of OU depth.
A slide showing a hierarchical diagram titled "Designing an Effective OU Structure" with an "Organization Root" at the top branching to four OUs: PCI, HIPAA, Public, and Internal. Each OU is connected downward to a corresponding account (e.g., PCI Account 01, HIPAA Account 01, etc.).
SCPs: how they work
  • SCPs are JSON policy documents attached to the root, an OU, or individual accounts.
  • SCPs set the maximum available permissions for principals in an account.
  • SCPs do not grant permissions; they only allow or deny actions.
  • Deny statements in SCPs override allow statements elsewhere.
  • SCPs are inherited by child OUs and accounts where attached (unless detached).
Effective permissions for a principal = intersection of allowed actions from IAM policies and the restrictions imposed by SCPs (and permission boundaries).
A slide titled "SCPs – Implementing Organization-Wide Guardrails" showing a vertical flowchart from "Organization SCPs" to "OU SCPs", "Account SCPs", "IAM Policies" and ending with "Effective Permissions = SCPs → IAM Policies." The image is credited to KodeKloud.
Example: deny RunInstances unless instance type is approved This SCP denies ec2:RunInstances unless the instance type is one of the approved list (valid JSON):
Common SCP patterns
  • Region restrictions (data residency): deny actions outside allowed regions.
  • Service restrictions: deny use of unapproved services.
  • Resource protection: prevent deletion/modification of critical security resources (CloudTrail, Config rules, GuardDuty).
  • Administrative boundaries: block privilege escalation by restricting IAM changes.
You can use the table below as a quick reference for common patterns and example intents.
A presentation slide titled "Common SCP Patterns – Implementing Security Guardrails with SCPs" showing a blue "Service Restrictions" box linked to two smaller boxes labeled "Prevent use of unapproved services" and "Block risky service actions." The slide also shows a © Copyright KodeKloud notice.
A slide titled "Common SCP Patterns – Implementing Security Guardrails with SCPs" showing a "Resource Protection" header. It lists three guardrails: prevent deletion of security resources, protect CloudTrail/Config/GuardDuty, and enforce encryption requirements.
Consolidated SCP example (multiple guardrails) A multi-statement SCP that combines region restrictions, service restrictions, protection of security resources, and prevention of privilege escalation:
Test SCPs thoroughly in non-production OUs before applying them broadly. Misapplied SCPs can block administrators and critical automation.
Cross-account access patterns
  • Best practice: avoid managing long-lived IAM users across accounts. Instead, use role assumption via STS:AssumeRole.
  • A federated user, CI/CD pipeline, or third-party integration calls STS:AssumeRole against a role in the target account. The role’s trust policy controls who can assume it (and may require an ExternalId).
  • Trust policies, permissions, and proper principals are the common failure points when cross‑account access fails.
A presentation slide titled "Cross–Account Access Patterns" showing a flow diagram: User/Identity → Assume Role Request (with ARN, External ID) → Target Role (trust policy with External ID) → Temporary Security Credentials.
Identity federation and SSO
  • Most organizations use identity federation (SSO) instead of long-lived IAM users.
  • IAM Identity Center (formerly AWS SSO) is common; enterprises also use identity providers like Okta or PingFederate.
  • Federated access centralizes authentication, improves attribution, and scales better—provided permissions mapping remains least privilege.
A presentation slide titled "Centralized Identity Management" showing a four-column comparison table (Model, Description, Pros, Cons) that outlines IAM centralization approaches like identity accounts, federated access, hybrid models and identity federation at scale.
Control Tower: bring up a secure landing zone AWS Control Tower automates a best‑practice landing zone with Organizations:
  • Account Factory for automated account provisioning
  • Predefined OU structure and managed SCPs
  • Preventive and detective guardrails
  • Centralized logging and single sign-on
  • A compliance dashboard for visibility
Control Tower is a strong option when you expect to manage dozens or hundreds of accounts and want an opinionated baseline.
A screenshot of an "AWS Control Tower – Multi-Account Governance" guardrail compliance dashboard showing a table of accounts, organizational units, guardrails, compliance status (compliant/non-compliant), and last-checked timestamps. The table rows list example accounts like Account-Dev-001 and Account-Prod-005 with their respective guardrail checks.
A slide titled "AWS Control Tower – Multi-Account Governance" showing an account provisioning workflow. It has a horizontal flowchart of five steps: Request New Account via Account Factory → Validate Request and Approvals → Provision Account & Configure Landing Zone → Apply Preventive & Detective Guardrails → Monitor Compliance via Dashboard.
IAM Access Analyzer and findings
  • IAM Access Analyzer detects unintended public or cross-account access.
  • Findings can be rated by severity and sent to Security Hub or routed to automated remediation via EventBridge (CloudWatch Events).
  • Common findings: public S3 buckets, cross-account role access, or IAM identities with external access.
A horizontal five-step flowchart titled "IAM Access Analyzer for Multi‑Account" showing a remediation workflow for cross‑account access findings. The steps run from "Detect Finding" to "Analyse Severity & Categorize," "Automated Finding Triage," "Integrate with Security Response Workflows," and "Continuous Validation & Policy Refinement."
A slide titled "IAM Access Analyzer for Multi-account" showing a table of access findings across several accounts. Columns list Resource, Finding Type, Severity, First Detected date, and Status (e.g., Open, In Progress, Closed).
EventBridge as a centralized event bus Use EventBridge to centralize events from member accounts and drive automations:
  • Route events to Lambda, Step Functions, CloudFormation StackSets, Config rules, Systems Manager Automation, Security Hub, SNS, etc.
  • Implement automated security governance patterns: detection → centralized event routing → responders → remediation and notifications.
Typical pattern:
  • Member accounts emit events → centralized EventBridge receives them → rules trigger responders (Lambda/SSM/StackSets) → remediation actions or notifications executed.
A diagram titled "Automated Security Governance" showing Account 01 and Account 02 sending events to Amazon EventBridge, which feeds a Security Automation workflow. The automation invokes AWS Lambda and a central EventBridge and integrates with AWS CloudFormation, AWS Config, and AWS Security Hub.
A diagram titled "Automating Security Controls Across Accounts" showing Amazon EventBridge triggering an AWS Lambda that runs security/compliance checks across Account A and Account B, then sends findings to AWS Systems Manager for centralized automation and remediation.
Monitoring and auditing at scale Centralize logs and findings with:
  • AWS CloudTrail (API activity)
  • AWS Config and Config aggregators (resource inventory & compliance)
  • AWS Security Hub (centralized findings)
  • Other telemetry services as needed
Collect telemetry centrally for trend analysis, alerting, and automated triage.
A diagram titled "Monitoring and Auditing At Scale" showing AWS Organizations feeding into Centralized Logging, Compliance Monitoring, and Operational Monitoring. On the left is a dashboard mockup with a security score, open findings graph, compliance metrics, and failed-checks bar charts.
Putting it all together — practical checklist
  1. Define account structure and OU design aligned with business drivers.
  2. Select baseline security controls (preventive via SCPs and detective via Config/CloudTrail/Security Hub).
  3. Choose a centralized identity model (SSO/federation or identity accounts) and implement least privilege.
  4. Automate continuous compliance and remediation with event-driven patterns (EventBridge → Lambda/SSM/StackSets → Security Hub/Config).
  5. Monitor, triage findings, and iterate on policies and automation.
A slide titled "Implementing a Secure Multi‑account Strategy" showing AWS Organizations at the center connected to Service Control Policies (SCPs) and an IAM key for cross‑account access. On the right, monitoring/aggregation services (AWS CloudTrail, AWS AppConfig, and AWS Security Hub) are linked for log and metric collection.
Common troubleshooting checklist
  • SCP not taking effect: verify where the SCP is attached (root/OU/account) and confirm inheritance.
  • Cross‑account access failures: check the role trust policy, ExternalId usage, and the principal’s permissions.
  • Inconsistent enforcement: confirm OU hierarchy and SCP inheritance.
  • Prevention guardrails bypassed: re-evaluate SCP scope and test coverage.
  • Use CloudTrail decision logs to investigate why an action was denied.
You can use this quick reference table for troubleshooting common issues:
A presentation slide titled "Common Multi‑account Security Challenges" showing a three‑column table of Challenge, Possible Cause, and Solution. Example rows mention issues like SCP not taking effect, cross‑account access failures, inconsistent policy enforcement and corresponding causes and fixes.
Final notes and references
  • Start with a clear OU and account design, apply guardrails via SCPs and account-level IAM policies.
  • Centralize telemetry (CloudTrail, AWS Config, Security Hub) and use EventBridge-driven automation for continuous governance and remediation.
  • Test policies in safe environments and roll them out incrementally.
  • Favor identity federation/SSO for scalability and auditability, and enforce least privilege.
Further reading: That’s it for this lesson. See you in the next one.

Watch Video