Cloud Security Reference Patterns: [Organisation or platform name]
Catalogue of approved cloud security patterns for identity, network,
data protection, workload protection, logging and detection, and
posture management, with the implementation on AWS, Microsoft Azure
and Google Cloud and a control mapping. It draws on the provider
security reference guidance: the AWS Security Reference Architecture
(AWS SRA, AWS Prescriptive Guidance, updated June 2026), the Microsoft
cloud security benchmark (MCSB v2, in preview, overview dated 15
January 2026, superseding v1) and the Google Cloud enterprise foundations blueprint
(formerly the security foundations blueprint). Controls map to the CSA
Cloud Controls Matrix v4.1 (January 2026), NIST SP 800-53 Rev. 5 and
CIS Benchmarks by topic. Replace every [bracketed] value and every
row marked Example.
Document control
| Field | Value |
|---|
| Document ID | [CSP-001] |
| Providers covered | [AWS / Microsoft Azure / Google Cloud] |
| Service models covered | [IaaS / PaaS / SaaS / Containers / Serverless] |
| Control frameworks mapped | CSA CCM v4.1; NIST SP 800-53 Rev. 5 [release 5.2.0]; [CIS Benchmark per provider, version]; [internal policy] |
| Version | [0.1] |
| Status | [Draft / In review / Approved / Superseded] |
| Owner | [@cloud-security-architect] |
| Approvers/reviewers | [@ciso-delegate], [@platform-lead], [@network-lead], [@security-operations-lead] |
| Classification | [Internal] |
| Last updated | YYYY-MM-DD |
| Provider documentation checked on | YYYY-MM-DD |
Approvals
| Role | Name | Decision | Date |
|---|
| Security architecture authority | [@name] | [Approve / Approve with conditions / Reject] | YYYY-MM-DD |
| Cloud platform owner | [@name] | [Decision] | YYYY-MM-DD |
| Security operations | [@name] | [Decision] | YYYY-MM-DD |
| Risk and compliance | [@name] | [Decision] | YYYY-MM-DD |
Revision history
| Version | Date | Author | Change summary |
|---|
| 0.1 | YYYY-MM-DD | [@name] | First draft |
| 1.0 | YYYY-MM-DD | [@name] | Example: Approved with 18 patterns across three providers |
| 1.1 | YYYY-MM-DD | [@name] | Example: Mapping moved from CCM v4.0 to v4.1; logging pattern updated after AWS SRA June 2026 revision |
Contents
1. Scope and how to use the patterns
1.1 Scope
| In scope | Out of scope |
|---|
| Example: All workloads in the organisation's AWS, Azure and Google Cloud tenants | SaaS applications assessed through supplier security assessment |
| Example: Platform shared services (identity federation, hub networks, logging, key management) | On-premises data centre controls |
| Example: Container platforms and serverless functions | Endpoint and end-user device security |
| [Item] | [Item] |
1.2 Pattern structure
Every pattern in sections 5 to 10 uses the same fields so that workload
teams can compare them and auditors can trace them.
| Field | Content |
|---|
| ID and name | PAT-[DOMAIN]-NNN, a short imperative name |
| Problem | The risk or recurring design question the pattern answers |
| Solution | The provider-neutral design, in three to six sentences |
| Provider implementation | The services and settings on each provider |
| Guardrails | Preventive and detective controls that enforce the pattern (IDs from the landing zone design) |
| Verification | How adoption is checked automatically and how often |
| Anti-patterns | Designs that look similar but are not acceptable |
| Mapping | CCM v4.1 domain, MCSB domain, SP 800-53 controls (section 12) |
| Status | Proposed, Approved, Deprecated or Retired (section 13) |
1.3 Adoption rule
Example: Production workloads adopt every pattern with status
Approved unless an exception is recorded in section 13. Non-production
workloads adopt the identity, logging and data protection patterns and
may defer the others. Pattern adoption is checked at the architecture
review (architecture/architecture-review-request) and continuously by
posture management (section 10).
2. Reference sources and versions
| Source | Publisher | Version or date used | What it contributes |
|---|
| AWS Security Reference Architecture (AWS SRA) | Amazon Web Services, AWS Prescriptive Guidance | Example: Revision of 30 June 2026 (major update; restructured December 2025 with separate deep-dive guides for identity, perimeter, forensics, generative AI and IoT) | Organisation and account structure, placement of security services |
| Microsoft cloud security benchmark | Microsoft | Example: v2 (preview), overview dated 15 January 2026; v1 still referenced by many tools | Control domains with Azure implementation and policy mappings |
| Enterprise foundations blueprint | Google Cloud | Example: Page last updated 15 May 2025; Terraform in terraform-example-foundation | Hierarchy, identity, network, detective and preventive control decisions |
| Cloud Controls Matrix | Cloud Security Alliance | v4.1 (released January 2026, 17 domains) | Provider-neutral control framework for mapping |
| NIST SP 800-53 Rev. 5 | NIST | Release [5.2.0, August 2025] | Control catalogue for mapping |
| CIS Benchmarks | Center for Internet Security | [Benchmark name and version per provider] | Configuration baseline checks |
| [Internal security policy] | [Owner] | [Version] | Mandatory requirements |
CSA has published a transition plan for STAR submissions: v4.0 and v4.1
are both accepted from March 2026, only v4.1 from December 2027, and the
v4.0.x CCM and CAIQ are withdrawn in January 2028. Check the current
timeline at cloudsecurityalliance.org.
3. Shared responsibility
| Layer | IaaS (virtual machines) | PaaS (managed database, app service) | Containers (managed Kubernetes) | Serverless functions | SaaS |
|---|
| Physical facilities and hardware | Provider | Provider | Provider | Provider | Provider |
| Hypervisor and host OS | Provider | Provider | Provider | Provider | Provider |
| Guest OS and runtime patching | Workload team | Provider | Shared: provider control plane, platform team node images | Provider | Provider |
| Network exposure and segmentation | Workload team within platform guardrails | Workload team (private endpoints, firewall rules) | Platform team (cluster), workload team (policies) | Workload team | Provider, customer configures access |
| Identity and access configuration | Customer | Customer | Customer | Customer | Customer |
| Data classification, encryption keys, backups | Customer | Customer | Customer | Customer | Customer, within provider features |
| Logging configuration and retention | Customer | Customer | Customer | Customer | Customer, within provider features |
Rows marked "Customer" are split between the platform team (shared
services and guardrails) and the workload team (resource configuration).
Record the split for each pattern in the RACI below.
| Activity | Platform team | Workload team | Security operations | Provider |
|---|
| Define and maintain patterns | A | C | C | I |
| Implement guardrails and shared services | R/A | I | C | I |
| Configure workload resources to the pattern | C | R/A | I | I |
| Monitor posture and triage findings | C | R | A | I |
| Respond to incidents | C | R | A | C (provider-side incidents) |
4. Security organisation structure
The AWS SRA organises a multi-account environment as: the Org
Management account; a Security OU with a Security Tooling account and a
Log Archive account; an Infrastructure OU with a Network account and a
Shared Services account; and a Workloads OU with application accounts.
The current SRA (revised 30 June 2026) also recommends a dedicated Monitoring account in
the Security OU as the home of organisation-wide CloudWatch telemetry,
with the Log Archive account kept as the immutable storage sink.
Azure and Google Cloud achieve the same separation with management
groups and subscriptions, and with folders and projects.
| Function | AWS (SRA) | Microsoft Azure | Google Cloud |
|---|
| Organisation administration | Org Management account | Tenant root and platform management group | Organization node, bootstrap and common folders |
| Security tooling, delegated administration | Security OU, Security Tooling account | Management or security subscription | Security project (Security Command Center administration) |
| Central log store | Security OU, Log Archive account (immutable sink) | Log Analytics workspace in the management subscription | Central logging project with aggregated sinks |
| Central telemetry analytics | Security OU, Monitoring account (CloudWatch delegated administrator) | Microsoft Sentinel workspace | Security Operations instance or log analytics project |
| Shared network | Infrastructure OU, Network account | Connectivity subscription (hub) | Shared VPC host projects per environment |
| Shared services (directory, key and secret services) | Infrastructure OU, Shared Services account | Identity and management subscriptions | Common folder projects (for example central Secret Manager and Cloud KMS projects) |
| Workloads | Workloads OU, application accounts | Landing zone subscriptions | Environment folders, workload projects |
5. Identity and access patterns
| ID | Pattern | Problem | Solution | AWS | Azure | Google Cloud |
|---|
| PAT-IAM-001 (Example) | Federate all human access | Local cloud users drift out of joiner-mover-leaver control | One authoritative identity provider, groups mapped to roles, no local users except break-glass | IAM Identity Center with external IdP | Microsoft Entra ID as the tenant directory | Cloud Identity or Workspace synced from the IdP |
| PAT-IAM-002 (Example) | Just-in-time privileged access | Standing administrator rights are the most valuable target | Eligible roles activated with approval, justification and a time limit | Identity Center permission sets with a temporary elevation workflow | Privileged Identity Management | Privileged Access Manager |
| PAT-IAM-003 (Example) | Keyless workload identity | Static keys leak through code and logs | Workloads and pipelines obtain short-lived tokens through federation | IAM roles, OIDC federation for pipelines | Managed identities, workload identity federation | Service account impersonation, Workload Identity Federation |
| PAT-IAM-004 (Example) | Protect the root of the organisation | Root or global administrator compromise defeats every other control | Centralise or remove root credentials, phishing-resistant MFA, alert on any use | Centralised root access management, SCP restrictions | Emergency access accounts, Conditional Access, alerting | Super administrator accounts limited and monitored |
| PAT-IAM-005 (Example) | Organisation-wide permission boundaries | One mis-scoped role can expose data to other tenants or the internet | Preventive policies at the hierarchy deny actions no workload needs | Service control policies, resource control policies | Azure Policy deny, management group RBAC | Organization Policy constraints, IAM deny policies |
| [PAT-IAM-NNN] | [Pattern] | [Problem] | [Solution] | [AWS] | [Azure] | [Google Cloud] |
Verification for every identity pattern: posture checks for local users,
keys older than [90 days], roles assigned to individuals, and use of
break-glass identities. Detailed design belongs in the IAM solution
design (security/iam-solution-design).
6. Network security patterns
| ID | Pattern | Problem | Solution | AWS | Azure | Google Cloud |
|---|
| PAT-NET-001 (Example) | Central egress inspection | Malware and data theft use outbound internet paths | All internet-bound traffic leaves through an inspected hub with domain filtering | Network Firewall in the Network account, Route 53 Resolver DNS Firewall | Azure Firewall in the hub, DNS proxy | Cloud NGFW, Secure Web Proxy, Cloud DNS policies |
| PAT-NET-002 (Example) | Private access to platform services | Managed services reachable from the internet | Private endpoints for platform services; public endpoints denied by policy | Interface VPC endpoints (PrivateLink), endpoint policies | Private endpoints, public network access disabled | Private Service Connect, VPC Service Controls perimeters |
| PAT-NET-003 (Example) | Edge protection for public applications | Web attacks and volumetric DDoS against public endpoints | Web application firewall and DDoS protection in front of every public endpoint | AWS WAF, Shield Advanced, CloudFront | Front Door or Application Gateway with WAF, DDoS Protection | Cloud Armor with external load balancing |
| PAT-NET-004 (Example) | Segmentation between workloads | Lateral movement across flat networks | Spoke per workload, deny spoke-to-spoke by default, identity-aware service-to-service access | VPC per account, Transit Gateway route tables, VPC Lattice auth policies | Spoke virtual networks, network security groups, hub routing | Separate VPCs or Shared VPC subnets, firewall policies with tags |
| PAT-NET-005 (Example) | Identity-aware access for administrators | VPNs grant broad network access | Application-level access proxy evaluates identity and device for each request | Verified Access | Entra Private Access, Conditional Access | Identity-Aware Proxy |
| [PAT-NET-NNN] | [Pattern] | [Problem] | [Solution] | [AWS] | [Azure] | [Google Cloud] |
Topology, address plan and hybrid connectivity are designed in the
Network architecture design (architecture/network-architecture-design).
7. Data protection and key management patterns
| ID | Pattern | Problem | Solution | AWS | Azure | Google Cloud |
|---|
| PAT-DAT-001 (Example) | Customer-managed keys for confidential data | Provider-managed keys do not separate duties for regulated data | Keys in a central key service, key administrators separate from data administrators, rotation yearly | KMS customer managed keys, key policies; CloudHSM where required | Key Vault or Managed HSM, customer-managed keys per service | Cloud KMS, Cloud HSM; Cloud EKM for externally held keys |
| PAT-DAT-002 (Example) | Secrets in a managed vault | Credentials in code, images and environment files | All secrets in a vault, retrieved at runtime by workload identity, automatic rotation | Secrets Manager | Key Vault secrets | Secret Manager |
| PAT-DAT-003 (Example) | Sensitive data discovery | Unknown copies of personal data in storage | Scheduled discovery scans, findings tagged with classification | Macie | Microsoft Purview, Defender for Cloud data security posture | Sensitive Data Protection |
| PAT-DAT-004 (Example) | Immutable, isolated backups | Ransomware and insiders delete backups with production rights | Backups in a separate container with separate identities, immutability lock, restore tests | AWS Backup vault lock, logically air-gapped vault | Azure Backup immutable vaults, soft delete, multi-user authorisation | Backup and DR Service backup vaults with enforced retention |
| PAT-DAT-005 (Example) | Encryption in transit everywhere | Plain-text traffic inside the network | TLS 1.2 or later on every endpoint, private certificates from a managed private CA | ACM, AWS Private CA | Key Vault certificates, App Gateway TLS policies | Certificate Manager, Certificate Authority Service |
| [PAT-DAT-NNN] | [Pattern] | [Problem] | [Solution] | [AWS] | [Azure] | [Google Cloud] |
Certificate hierarchy and lifecycle are designed in the PKI and
certificate architecture (security/pki-certificate-architecture).
Residency constraints are assessed in the Data residency and sovereignty
assessment (compliance/data-residency-sovereignty-assessment).
8. Workload protection patterns
| ID | Pattern | Problem | Solution | AWS | Azure | Google Cloud |
|---|
| PAT-WKL-001 (Example) | Hardened, rebuilt images | Configuration drift and unpatched long-running hosts | Golden images built by pipeline against a CIS Benchmark profile, rebuilt at least monthly | EC2 Image Builder, Systems Manager patching | Azure Image Builder, Update Manager | OS images built by Cloud Build, VM Manager patching |
| PAT-WKL-002 (Example) | Continuous vulnerability scanning | Known vulnerabilities in images and running workloads | Scan registries and running workloads; fix by severity service levels | Amazon Inspector | Defender for Cloud (servers and containers plans) | Artifact Analysis, Security Command Center vulnerability findings |
| PAT-WKL-003 (Example) | Signed artefacts only | Tampered images deployed from a compromised pipeline | Images signed in the pipeline, admission control verifies signature and provenance | Signer, EKS admission policies | Container registry signing with Notation, AKS policy | Binary Authorization |
| PAT-WKL-004 (Example) | Runtime threat detection | Attacks inside running workloads go unnoticed | Runtime agents or managed detection on hosts, containers and functions | GuardDuty runtime and malware protection | Defender for Cloud workload protection | Security Command Center threat detection services |
| PAT-WKL-005 (Example) | Confidential processing for high-sensitivity data | Data in use exposed to host operators | Isolated enclaves or confidential VMs for designated workloads | Nitro Enclaves | Confidential VMs and containers | Confidential VMs and Confidential GKE nodes |
| [PAT-WKL-NNN] | [Pattern] | [Problem] | [Solution] | [AWS] | [Azure] | [Google Cloud] |
Supply chain controls for the pipeline itself are covered in Supply
chain security (security/supply-chain-security).
9. Logging, detection and response patterns
| ID | Pattern | Problem | Solution | AWS | Azure | Google Cloud |
|---|
| PAT-LOG-001 (Example) | Organisation-wide audit trail | Missing or deletable control-plane logs | Organisation-level audit logging to the log archive, immutable, workload teams read only | Organisation CloudTrail trail to the Log Archive account | Activity log and Entra logs to the central workspace via diagnostic settings | Organisation aggregated sinks to the central logging project, locked buckets |
| PAT-LOG-002 (Example) | Central security data store | Investigations need data from many accounts and tools | Normalised security data in one store with long retention | Example: CloudWatch Unified Data Experience in the Monitoring account as primary, Amazon Security Lake as secondary, per the June 2026 SRA revision | Microsoft Sentinel workspace | Google Security Operations or BigQuery log analytics |
| PAT-LOG-003 (Example) | Managed threat detection everywhere | Regions or accounts left without detection | Detection enabled in every region and container by delegated administration | GuardDuty with delegated administrator | Defender for Cloud plans on all subscriptions | Security Command Center at organisation level |
| PAT-LOG-004 (Example) | Aggregated findings and automated response | Findings scattered across consoles, slow response | Findings aggregated, triaged by severity and routed to ticketing; safe automated containment | Security Hub, EventBridge, Security Incident Response | Defender XDR, Sentinel automation rules and playbooks | Security Command Center, Security Operations playbooks |
| [PAT-LOG-NNN] | [Pattern] | [Problem] | [Solution] | [AWS] | [Azure] | [Google Cloud] |
Incident handling follows the Incident response plan
(compliance/incident-response-plan); alert definitions live in the
Observability specification (ops/observability-spec).
10. Posture management and policy as code
| ID | Pattern | Solution | AWS | Azure | Google Cloud |
|---|
| PAT-POS-001 (Example) | Continuous posture assessment | Configuration recorded and assessed against the CIS Benchmark and internal rules in every container | AWS Config, Security Hub CSPM standards | Defender for Cloud CSPM, MCSB initiative in Azure Policy | Security Command Center posture management |
| PAT-POS-002 (Example) | Pre-deployment policy checks | IaC plans evaluated against the same rules before apply | Policy as code in the pipeline, CloudFormation Guard or open policy engines | Azure Policy, pipeline checks on Bicep or Terraform | Policy checks on Terraform plans, Organization Policy dry-run |
| PAT-POS-003 (Example) | External exposure review | Unintended public or cross-account access | Automated analysis of resource policies and network reachability | IAM Access Analyzer, Network Access Analyzer | Defender for Cloud attack path analysis |
| [PAT-POS-NNN] | [Pattern] | [Solution] | [AWS] | [Azure] | [Google Cloud] |
Remediation service levels (Example): critical findings fixed within
[2 days], high within [7 days], medium within [30 days]; anything
longer needs an exception in section 13.
11. Provider equivalence table
| Capability | AWS | Microsoft Azure | Google Cloud |
|---|
| Hierarchy and isolation unit | Organizations, OU, account | Management group, subscription | Folder, project |
| Preventive organisation policy | SCP, RCP, declarative policies | Azure Policy (deny) | Organization Policy |
| Human identity | IAM Identity Center | Microsoft Entra ID | Cloud Identity |
| Just-in-time privilege | Temporary elevated access workflow | Privileged Identity Management | Privileged Access Manager |
| Workload identity | IAM roles | Managed identities | Service accounts with Workload Identity Federation |
| Key management | KMS, CloudHSM | Key Vault, Managed HSM | Cloud KMS, Cloud HSM, Cloud EKM |
| Secrets | Secrets Manager | Key Vault | Secret Manager |
| Network firewall | Network Firewall | Azure Firewall | Cloud NGFW |
| Data exfiltration perimeter | Endpoint and resource control policies | Private endpoints with network policies | VPC Service Controls |
| Web and DDoS protection | AWS WAF, Shield Advanced | Azure WAF, DDoS Protection | Cloud Armor |
| Threat detection | GuardDuty | Defender for Cloud | Security Command Center |
| Posture management | Security Hub CSPM, Config | Defender for Cloud CSPM | Security Command Center |
| SIEM and security data | CloudWatch Unified Data Experience, Security Lake, partner SIEM | Microsoft Sentinel | Google Security Operations |
| Private CA | AWS Private CA | Microsoft Cloud PKI (Intune, device certificates), or a partner or self-managed CA with Key Vault | Certificate Authority Service |
12. Control mapping
CSA CCM v4.1 domains: A&A Audit and Assurance; AIS Application and
Interface Security; BCR Business Continuity Management and Operational
Resilience; CCC Change Control and Configuration Management; CEK
Cryptography, Encryption and Key Management; DCS Datacenter Security;
DSP Data Security and Privacy; GRC Governance, Risk Management and
Compliance; HRS
Human Resources Security; IAM Identity and Access Management; IPY
Interoperability and Portability; IVS Infrastructure and Virtualization
Security; LOG Logging and Monitoring; SEF Security Incident Management,
E-Discovery and Cloud Forensics; STA Supply Chain Management,
Transparency and Accountability; TVM Threat and Vulnerability
Management; UEM Universal Endpoint Management.
MCSB v2 (preview) domains: NS Network Security, IM Identity Management,
PA Privileged Access, DP Data Protection, AM Asset Management, LT
Logging and Threat Detection, IR Incident Response, PV Posture and
Vulnerability Management, ES Endpoint Security, BR Backup and Recovery,
DS DevOps Security, AI Artificial Intelligence Security. MCSB v1 also
has GS Governance and Strategy.
| Pattern | CCM v4.1 domain | MCSB domain | SP 800-53 Rev. 5 | CIS Benchmark topic | Evidence |
|---|
| PAT-IAM-001 (Example) | IAM | IM | AC-2, IA-2, IA-8 | Identity and access management | IdP configuration export, posture check for local users |
| PAT-IAM-002 (Example) | IAM | PA | AC-2(6), AC-6 | Identity and access management | Elevation logs |
| PAT-NET-002 (Example) | IVS | NS | SC-7, AC-4 | Networking | Policy assignment report |
| PAT-DAT-001 (Example) | CEK | DP | SC-12, SC-28 | Storage and database | Key policy review |
| PAT-DAT-004 (Example) | BCR | BR | CP-9, CP-10 | Storage | Restore test reports |
| PAT-WKL-002 (Example) | TVM | PV | RA-5, SI-2 | Compute | Scan coverage dashboard |
| PAT-LOG-001 (Example) | LOG | LT | AU-2, AU-9, AU-11 | Logging | Trail and sink configuration |
| PAT-LOG-004 (Example) | SEF | IR | IR-4, IR-5 | Monitoring | Ticket metrics |
| PAT-POS-001 (Example) | CCC, TVM | PV | CM-6, CA-7 | All sections | Benchmark pass rate |
| [PAT-XXX-NNN] | [Domain] | [Domain] | [Controls] | [Topic] | [Evidence] |
13. Exceptions and pattern lifecycle
13.1 Pattern lifecycle
| Status | Meaning | New adoption allowed |
|---|
| Proposed | Under review, may be piloted | Pilot only |
| Approved | Mandatory or recommended per section 1.3 | Yes |
| Deprecated | Replaced; existing adopters keep it until the migration date | No |
| Retired | No longer supported | No |
13.2 Exceptions
| ID | Workload | Pattern | Deviation | Compensating control | Risk owner | Expiry |
|---|
| EXC-001 (Example) | Legacy reporting VM | PAT-WKL-001 | Image cannot be rebuilt monthly due to vendor support terms | Host firewall allow-list, weekly vulnerability scan | @reporting-owner | YYYY-MM-DD |
| EXC-002 (Example) | Partner SFTP | PAT-NET-002 | Public endpoint required | Source IP allow-list, WAF, dedicated container | @integration-lead | YYYY-MM-DD |
| [EXC-NNN] | [Workload] | [Pattern] | [Deviation] | [Control] | [@owner] | YYYY-MM-DD |
Expired exceptions are reported to the risk owner and the security
architecture authority each month. Record accepted risk in the Risk
register (compliance/risk-register).
14. Adoption register and traceability
| Workload | Provider | PAT-IAM | PAT-NET | PAT-DAT | PAT-WKL | PAT-LOG | PAT-POS | Exceptions | Last verified |
|---|
| Payments API (Example) | AWS | 5/5 | 4/5 | 5/5 | 4/5 | 4/4 | 3/3 | none | YYYY-MM-DD |
| Customer portal (Example) | Azure | 5/5 | 5/5 | 4/5 | 5/5 | 4/4 | 3/3 | none | YYYY-MM-DD |
| Analytics platform (Example) | Google Cloud | 4/5 | 3/5 | 5/5 | 3/5 | 4/4 | 3/3 | EXC-NNN | YYYY-MM-DD |
| [Workload] | [Provider] | [n/N] | [n/N] | [n/N] | [n/N] | [n/N] | [n/N] | [EXC IDs] | YYYY-MM-DD |
Traceability runs from requirement to pattern, guardrail, control and
evidence:
| Requirement | Pattern | Guardrail (landing zone) | Control | Evidence |
|---|
| Example: Policy 4.2, no standing admin rights | PAT-IAM-002 | GR-NNN | AC-2(6) | Elevation log export |
| [Requirement] | [Pattern] | [GR-NNN] | [Control] | [Evidence] |
15. Glossary and references
15.1 Glossary
| Term | Meaning |
|---|
| Pattern | An approved, reusable design for a recurring security problem |
| Guardrail | An automated preventive or detective control applied across many containers |
| Delegated administrator | A security account or subscription allowed to manage a security service for the whole organisation |
| CSPM | Cloud security posture management: continuous assessment of configuration against rules |
| Data perimeter | Controls that keep data inside trusted identities, resources and networks |
| Break-glass | Emergency access used only when normal access fails |
15.2 References
| Reference | Version or date checked | Use |
|---|
| AWS Security Reference Architecture, docs.aws.amazon.com/prescriptive-guidance | Revision 30 June 2026, checked YYYY-MM-DD | Section 4, AWS columns |
| Microsoft cloud security benchmark, learn.microsoft.com/security/benchmark | v2 preview (overview dated 15 January 2026) | Domains, Azure columns |
| Google Cloud enterprise foundations blueprint, docs.cloud.google.com/architecture/security-foundations | Updated 15 May 2025 | Google columns |
| CSA Cloud Controls Matrix, cloudsecurityalliance.org | v4.1, January 2026 | Section 12 |
| NIST SP 800-53 Rev. 5, csrc.nist.gov | Release 5.2.0, August 2025 | Section 12 |
| CIS Benchmarks, cisecurity.org | [Version per provider] | Section 10 |
Landing zone design (architecture/landing-zone-design) | [Version] | Guardrail IDs |
Standard and licence note
Structure aligned with the AWS Security Reference Architecture, the
Microsoft cloud security benchmark, the Google Cloud enterprise
foundations blueprint and the CSA Cloud Controls Matrix v4.1, with
control identifiers from NIST SP 800-53 Rev. 5. Only account, domain,
product and control names and identifiers are used, for mapping; no
recommendation, control or benchmark text is reproduced, and CIS
Benchmarks are referenced by name and topic only. The pattern text is
original and written for this collection. AWS, Microsoft Azure, Google
Cloud and their products are trademarks of their owners; this template
is not affiliated with or endorsed by them, the Cloud Security Alliance
or the Center for Internet Security. Check the current official
documentation and framework versions before use. Template licence: MIT.