Cloud Security Alliance CCSKv5: Current Cloud Security Domains
CCSK v5 is built for the way cloud security is actually practiced today. Cloud programs are no longer just virtual machines in somebody else’s data center. They include managed services, containers, serverless workloads, SaaS, CI/CD, identity federation, security telemetry, data lakes, Zero Trust, AI-enabled services, and complex third-party dependencies. The current Cloud Security Alliance exam reflects that wider operating model.
Cloud Security Alliance CCSKv5 supports the current version of the Certificate of Cloud Security Knowledge. CSA identifies v5 as the latest exam version and organizes it into twelve domains covering cloud architecture, governance, risk and compliance, organizational management, IAM, monitoring, infrastructure, workloads, data, application security, incident response and resilience, plus related technologies and strategies.
Start with service models, deployment models, virtualization, cloud-native architecture, control planes, management planes, APIs, and the shared-responsibility model.
The central exam skill is recognizing which party controls which layer. A provider may secure physical infrastructure while the customer still owns identity, network policy, application configuration, data access, and workload security.
Shared responsibility changes with the service model. IaaS usually leaves more operating-system, network, and workload configuration to the customer; managed platform services move more operational responsibility to the provider; SaaS shifts still more infrastructure responsibility while leaving the customer accountable for users, data, permissions, configuration, and usage. Candidates should reason from the actual service boundary instead of applying one memorized responsibility list to every cloud scenario.
Modern architecture also includes APIs, control planes, managed identities, serverless services, containers, and hybrid or multi-cloud connections. A secure design should show where trust changes, which identities can modify resources, and where logs or policy enforcement prove the intended architecture is operating.
The cloud security fundamentals is useful for reinforcing this architecture-first view.
Cloud governance connects business requirements to technical policy. It covers account structure, ownership, approved services, regions, logging, security baselines, exception handling, and change control.
Risk management asks what could affect objectives and which controls reduce that exposure. Compliance adds external or internal obligations that must be translated into technical evidence.
Organizational management matters because cloud security spans architecture, security, engineering, legal, procurement, operations, development, and business teams. A technically strong control can fail when ownership is unclear.
Governance should make those ownership boundaries explicit. Account or subscription structure, approved services, permitted regions, mandatory logging, encryption expectations, exception processes, and deployment standards should map to responsible teams and measurable controls.
Risk and compliance should then use evidence from the platform. A rule that says storage must not be public needs configuration visibility, prevention or alerting, an exception path, and remediation ownership. A compliance statement without technical evidence is not enough to demonstrate control effectiveness.
IAM is one of the most important cloud-security domains because cloud platforms are API-driven. Human users, administrators, workloads, automation, and third parties all authenticate to control-plane services.
Study federation, MFA, least privilege, role design, privileged access, workload identity, service accounts, access review, and credential lifecycle.
Cloud identity and access gives these topics practical structure. Human administrators, workloads, CI/CD pipelines, SaaS integrations, and serverless functions can all hold permissions, so identity architecture must cover both human and machine subjects.
Prefer short-lived, scoped workload identity over embedded static secrets when the platform supports it. Review roles and service accounts for unused privilege, and design federation so claims, groups, and sessions translate into intended cloud permissions rather than broad default access.
The Zero Trust provides useful context for explicit verification, contextual access, and limiting persistent trust.
Monitoring turns architecture into observable security. Identity logs, network flow, cloud audit events, workload telemetry, application logs, and data-access records help teams detect misuse and investigate incidents.
Collection needs a purpose. Teams should know which events matter, how long they are retained, how time and identity are normalized, and what response workflow follows a meaningful alert.
Monitoring also needs health checks. A silent logging failure can create a dangerous blind spot even when workloads remain online.
Security logging and telemetry should be designed around investigative questions. Identity changes, control-plane actions, network flows, workload events, and sensitive data access can all matter, but the organization also needs normalized time and identity context, retention, searchability, and ownership of alerts.
Monitor the telemetry pipeline itself. A disabled audit source, broken export, expired credential, or retention change can remove evidence silently. Coverage and freshness are part of the control environment.
Infrastructure security includes network segmentation, security groups, private connectivity, DNS, management access, hardening, and control-plane protection.
Workload security changes by platform. Virtual machines need operating-system controls; containers need image, registry, orchestration, runtime, and cluster controls; serverless services shift more infrastructure operation to the provider while leaving identity, code, dependencies, and data with the customer.
Data security spans classification, encryption, access, key management, masking, retention, backup, deletion, and monitoring. The secrets management is relevant because machine credentials often connect cloud workloads to data and services.
Infrastructure security also includes posture management: public exposure, overly broad security groups, missing logging, unencrypted storage, insecure administrative paths, and unsupported configurations should be visible and assigned for remediation. Cloud security posture management is the operational bridge between governance policy and the real configuration state of a large cloud estate.
Workload controls differ by compute model. Virtual machines still need operating-system and endpoint controls; containers add image, registry, orchestrator, and runtime concerns; serverless functions shift host management to the provider but leave code, identity, dependencies, events, and data permissions with the customer.
Data security and privacy should follow the lifecycle across storage, analytics, backups, logs, caches, and third parties. Encryption is necessary in many scenarios, but classification, access, key governance, masking, retention, auditability, and deletion determine whether the broader control objective is met.
Cloud applications inherit traditional software risks and add infrastructure-as-code, APIs, managed services, supply-chain dependencies, and automated deployment.
Secure development should include threat modeling, code review, dependency checks, secrets handling, environment separation, testing, pipeline controls, and deployment validation.
DevSecOps is not a separate product. It is the practice of placing appropriate security checks into the software-delivery lifecycle so teams can find and fix problems before they reach production.
DevSecOps can include threat modeling, dependency checks, secret scanning, infrastructure-as-code policy, image scanning, artifact integrity, environment separation, and deployment validation. The control should appear at the stage where it can prevent or detect the relevant defect efficiently.
CI/CD is also part of the attack surface. Repositories, runners, registries, signing systems, deployment tokens, and cloud service identities may all have production influence.
Cloud incidents can involve account takeover, exposed services, data leakage, malicious workloads, compromised automation, provider outage, or destructive administrative action.
Response teams need access to logs, snapshots, configuration history, identities, and recovery resources even when parts of the environment are impaired.
The architecture should also support recovery: backup, replication, tested failover, protected credentials, and clear ownership. Resilience is the ability to restore trusted service, not merely keep a resource running.
Incident response in cloud environments needs access to control-plane history, identity logs, snapshots, configuration state, and provider-specific evidence. Recovery should restore trusted identity, network, logging, and data state, not just recreate compute resources.
Practice failure modes beyond regional outage: compromised automation, accidental public exposure, destructive administration, lost credentials, or a disabled logging pipeline. The response plan should identify both the technical recovery path and the team that owns each dependency.
The final domain covers related technologies and strategies, which is where current cloud security connects with Zero Trust, AI, data lakes, automation, and other modern architectures. Candidates should expect these topics to interact with the core domains rather than appear as isolated definitions.
Related technologies should be evaluated through the same security model as the rest of the cloud. AI services introduce sensitive prompts, model or retrieval data, and new automation paths; data lakes concentrate large datasets and complicated access; Zero Trust changes how trust is evaluated; and automation can magnify both good policy and bad configuration. The candidate should ask which identity acts, which data is exposed, which control plane is involved, what telemetry proves behavior, and how the system can recover if the automation or service is compromised.
A useful study method is to choose one cloud application and walk through all twelve domains. Define service model, governance, risks, identities, monitoring, network, workload controls, data controls, application security, response, and recovery. Then introduce a failure such as credential compromise or public data exposure and follow the controls end to end.
CCSK v5 rewards connected reasoning. The strongest candidate can explain how cloud architecture, governance, identity, workload, data, software, monitoring, and resilience support one another instead of studying twelve unrelated lists.
CSA’s current CCSK v5 exam is open-book and online, with 60 multiple-choice questions, 120 minutes, an 80% passing score, and two attempts included with an exam purchase. Version 5 has been the current examination since July 16, 2024.
A strong final review should map one cloud application across all twelve domains and then introduce change: add serverless processing, move sensitive data into analytics, connect a third-party SaaS tool, or introduce generative AI. The candidate should be able to explain how responsibility, identity, telemetry, workload security, data controls, application security, governance, and recovery change together.
