Skip to main content
Checking session
← Back to SAA-C03

Certification

AWS Certified Solutions Architect – Associate

Design Secure Architectures

Exam domain

Tasks

  1. 1.1 Design secure access to AWS resources

    Secure access on AWS combines strong identity, least-privilege authorization, temporary credentials, and multi-account governance. Design centralized workforce access, explicit workload identities, separation of duties, and traceable access across accounts.

    Security best practices in IAM

    Core access concepts

    • Multi-account controls: Use AWS Organizations, organizational units (OUs), service control policies (SCPs), and cross-account roles to separate workloads and establish permission guardrails.
    • Workforce access: Use AWS IAM Identity Center with an external identity provider when appropriate, permission sets, and short-lived role credentials.
    • Workload access: Use IAM roles for AWS services and applications instead of distributing long-lived access keys.
    • Least privilege: Grant only required actions on required resources, add applicable conditions, and refine permissions using access evidence.
    • Shared responsibility: AWS secures the underlying cloud infrastructure; customers configure identities, permissions, data protection, and account controls.

    Root user and human access

    • Protect the root user, enable multi-factor authentication (MFA), avoid routine root use, and establish a monitored recovery or emergency-access process.
    • Prefer federation and IAM Identity Center for workforce access instead of creating individual IAM users.
    • Require MFA for privileged human access and use short-lived sessions appropriate to the task and risk.
    • Avoid long-lived access keys. Where a legacy exception exists, protect, monitor, rotate, and remove the key as soon as feasible.

    Flexible authorization

    • Identity-based policies: Attach permissions to IAM users, groups, or roles. Customer managed policies can be reused; inline policies remain attached to one identity.
    • Resource-based policies: Attach permissions to supported resources, including Amazon S3 buckets, AWS Key Management Service keys, Amazon Simple Queue Service queues, and AWS Lambda functions.
    • Conditions: Narrow authorization using supported context keys such as principal tags, MFA context, resource tags, source networks, VPC endpoints, and resource ARNs.
    • Permissions boundaries: Set the maximum permissions an identity-based policy can grant to an IAM user or role. A boundary grants no permission by itself, and resource-based policy interactions require evaluation of the specific principal and policy context.
    • Temporary credentials: Use AWS Security Token Service (AWS STS) role sessions and control session duration according to workload and user requirements.

    Cross-account access

    For cross-account role access, the target account role trust policy identifies who may assume the role, and permissions policies determine what the resulting role session can do. Both sides of the authorization relationship must be evaluated.

    • Map workforce groups to IAM Identity Center permission sets for consistent job-based access across accounts.
    • Use sts:AssumeRole and temporary credentials for cross-account operations.
    • Require MFA, approval, or additional conditions for sensitive roles where appropriate.
    • Create tightly controlled emergency access with strong monitoring and a tested recovery process.

    Multi-account governance

    • AWS Control Tower: Establishes and governs a landing zone and applies AWS Control Tower controls. The earlier term guardrails may appear in legacy material.
    • AWS Organizations and OUs: Group accounts according to workload, environment, business, and security boundaries.
    • SCPs: Limit available permissions for member accounts but do not grant permissions.
    • Delegated administration: Assign supported service administration to dedicated accounts while retaining centralized governance.

    Identity-based and resource-based policies

    Policy type Use Examples
    Identity-based Grant an IAM principal permissions to act on specified resources. Workforce roles, application roles, and service roles.
    Resource-based Grant specified principals access through a supported resource and express resource-specific constraints. S3 bucket policies, KMS key policies, SQS queue policies, and Lambda resource policies.

    For cross-account access, evaluate the identity policy, resource policy, trust policy, SCPs, permissions boundaries, session policies, and explicit denies that apply to the request.

    Federation and identity lifecycle

    • Use IAM Identity Center or supported SAML 2.0 and OpenID Connect federation to centralize workforce identity.
    • Map identity-provider groups and attributes to permission sets or roles according to job function.
    • Deprovisioning a user at the corporate identity provider prevents or removes future access according to provisioning and assignment behavior, but existing AWS sessions can remain usable until they expire or are otherwise revoked.
    • Coordinate identity provisioning, assignment removal, session duration, MFA, and logging to meet the required revocation objective.

    Design examples

    • Shared S3 data: Combine a bucket policy that identifies approved cross-account roles with restrictive role permissions, KMS key access when applicable, and relevant conditions.
    • Human access across accounts: Use IAM Identity Center permission sets, MFA, short sessions, and an elevation process for privileged access.
    • Production account controls: Apply SCPs and AWS Control Tower controls to the production OU, and delegate supported security-service administration to a security account.
    • Application credentials: Assign instance profiles, task roles, or function execution roles rather than embedding static credentials in hosts or images.
  2. 1.2 Design secure workloads and applications

    Secure workload design applies defense in depth across identity, network, application, data, detection, and external-connectivity layers. Select controls according to data classification, business and compliance requirements, service capabilities, and the threat model.

    Core security controls

    • Configuration and credentials: Store supported secrets in AWS Secrets Manager or sensitive configuration in AWS Systems Manager Parameter Store, with access and encryption controls appropriate to the data.
    • Private service access: Interface VPC endpoints use AWS PrivateLink for supported services. Amazon S3 and Amazon DynamoDB also support gateway VPC endpoints, which use route-table entries rather than AWS PrivateLink.
    • Traffic controls: Security groups provide stateful controls at elastic network interfaces; network ACLs provide stateless subnet-level controls.
    • Application identity: Amazon Cognito supports customer and application-user identity scenarios. IAM Identity Center supports workforce access to AWS accounts and applications.
    • Web and threat protection: AWS WAF filters web requests and common web exploits, AWS Shield provides DDoS protection, and Amazon GuardDuty detects supported threat activity. These services have different purposes and are not interchangeable.

    VPC security architecture

    • Subnets: Separate internet-facing entry points, application tiers, and data tiers according to routing and security requirements.
    • Route tables: Determine reachable network paths. Keep private resources without a direct route to an internet gateway when they do not require direct internet reachability.
    • Security groups: Permit only required sources, destinations, ports, and protocols.
    • Network ACLs: Add stateless subnet controls where the threat model or governance requirements justify them.
    • NAT gateways: Allow resources in private subnets to initiate supported outbound connections through a public NAT gateway. NAT does not support unsolicited inbound connections to those resources.

    Segmentation and defense in depth

    • Segment presentation, application, and data tiers by function and sensitivity.
    • Use security-group references to authorize required tier-to-tier paths without broad address ranges where supported.
    • Use AWS Transit Gateway, VPC peering, or other connectivity patterns according to topology, scale, transitivity, inspection, and routing requirements.
    • Use AWS Firewall Manager to manage supported security policies across accounts when centralized policy administration is required.

    Application and workload protection

    • Authentication: Use Amazon Cognito for applicable customer-facing application identity, or another identity solution that satisfies protocol and lifecycle requirements.
    • Threat detection: Use GuardDuty for supported threat findings and combine it with service logs, AWS CloudTrail, and response workflows.
    • Sensitive-data discovery: Use Amazon Macie to discover supported categories of sensitive data in Amazon S3.
    • Secrets: Secrets Manager can store, retrieve, and rotate supported secrets, but rotation requires a configured rotation function or supported managed rotation and appropriate permissions.
    • Encryption: Choose encryption at rest and in transit according to classification, requirements, service support, and threats. AWS Key Management Service and TLS are common controls, not universal instructions independent of context.
    • Web protection: Apply AWS WAF rules to supported web endpoints and choose the applicable AWS Shield tier based on DDoS risk and support requirements.

    External and hybrid connections

    • AWS Site-to-Site VPN: Provides IPsec connectivity over the network between a customer gateway and AWS.
    • AWS Direct Connect: Provides dedicated private connectivity, but traffic is not encrypted by default. Use supported options such as MAC Security or a VPN over Direct Connect when encryption requirements call for them.
    • VPC endpoints: Keep supported service traffic on the AWS network without requiring an internet gateway or NAT path.
    • Resilient hybrid design: Select redundant connections, locations, and routing according to availability and recovery requirements.

    Design examples

    • Three-tier workload: Place the load balancer in public subnets and application and database tiers in private subnets; authorize only required load-balancer-to-application and application-to-database traffic; retrieve configured secrets at runtime.
    • Secured API: Use Amazon API Gateway with the appropriate authorizer, AWS WAF for supported web-request filtering, and CloudTrail plus service logs and threat-detection findings for visibility.
    • Hybrid workload: Use redundant Direct Connect connections where dedicated connectivity is required, add explicit encryption when required, and use the correct interface or gateway endpoint type for private AWS service access.
  3. 1.3 Determine appropriate data security controls

    Data-security architecture combines access governance, encryption, classification, retention, backup, recovery, monitoring, and evidence. Controls must follow the data's classification, applicable requirements, service capabilities, recovery objectives, and threat model.

    Core data-security concepts

    • Access and governance: Determine who may access data, from which paths, and under what conditions using applicable identity, resource, network, and service controls.
    • Classification: Classify data by sensitivity, business value, residency, and regulatory obligations before selecting controls.
    • Recovery: Choose backup, replication, and restore mechanisms that satisfy recovery point objectives (RPOs) and recovery time objectives (RTOs).
    • Retention: Define retention and deletion according to stated business and legal requirements rather than a universal period.
    • Encryption and key management: Select service-managed or customer-controlled key options according to control, audit, availability, cost, and operational requirements.

    Encryption and key management

    • At rest: Amazon S3 encrypts new object uploads by default with server-side encryption using Amazon S3 managed keys (SSE-S3). Use a customer managed KMS key or another supported option when key-policy, audit, separation, or regulatory requirements call for it.
    • In transit: Use TLS and service-supported certificate mechanisms where transport protection is required. AWS Certificate Manager supplies certificates to supported integrated services such as load balancers and CloudFront; Amazon RDS DB endpoints use RDS-managed certificate authorities and client trust configuration.
    • KMS terminology: Use KMS key and customer managed KMS key. AWS managed keys and customer managed keys differ in who creates and manages their policies, lifecycle, and supported configuration.
    • Authorization: Evaluate KMS key policies, IAM policies, grants, service integration, and separation-of-duties requirements.
    • Rotation: Choose supported automatic or on-demand rotation according to KMS key type, key material origin, service capability, and policy requirements. Do not assume every key supports the same rotation mechanism or interval.
    • AWS CloudHSM and custom key stores: Consider them when requirements call for dedicated hardware security modules or specific control of key material. They add cost, availability-design, integration, and operational responsibilities and are not universally the best option.

    Lifecycle, retention, and classification

    • Use Amazon S3 Lifecycle rules to transition or expire objects according to access patterns and retention requirements, selecting the precise S3 storage class for the retrieval and durability need.
    • Use S3 Object Lock when a supported write-once-read-many retention control is required.
    • Apply data classification labels or tags through an approved governance process; do not assume discovery automatically establishes authoritative classification.
    • Use Amazon Macie for supported sensitive-data discovery in Amazon S3. Macie does not automatically classify arbitrary data across all AWS services.
    • Use AWS Lake Formation permissions and governance features for applicable data-lake architectures.

    Backup, replication, and recovery

    • AWS Backup: Centralizes policy-based backup for supported services and can support cross-account and cross-Region copies when configured.
    • Amazon S3 replication: Copies eligible objects according to configured same-Region or cross-Region replication rules; replication and backup are different controls.
    • Amazon RDS point-in-time recovery: Restores a DB instance to a supported point within the automated-backup retention period.
    • Amazon DynamoDB point-in-time recovery: Uses continuous backups to restore a table to a supported point in its recovery window.
    • Test restores and failover procedures to validate actual RPO, RTO, permissions, dependencies, and runbooks.

    Compliance evidence and security findings

    • AWS Artifact: Provides access to AWS compliance reports and supports management of applicable agreements; it does not automatically map each customer workload to a compliance framework.
    • AWS Config: Records supported resource configuration and evaluates configured rules or conformance packs.
    • AWS CloudTrail: Records supported AWS API and account activity and supplies evidence for investigation and audit workflows.
    • AWS Security Hub: Aggregates and normalizes supported security findings and supports posture-management and response workflows; it is not itself the authoritative activity log for all data access.

    Design examples

    • Sensitive objects in S3: Begin with S3 default encryption, then select a customer managed KMS key when the required key policy, audit, or separation controls justify it. Restrict access with IAM and bucket policies and monitor relevant activity.
    • Application-to-database transport: Configure the database and client to require TLS, trust the current RDS certificate authority bundle, and plan certificate rotation.
    • Recovery design: Configure service-specific backup and point-in-time recovery controls from stated RPO, RTO, retention, residency, and account-isolation requirements rather than applying generic retention numbers.

    Additional AWS references

Practice this domain

Try a Domain Sample

Use a focused public sample to validate your understanding before creating an account.

Try Domain Sample