STEP 1 · Study

SAP-C02 StudySolutions Architect – Professional

Read the key points for each exam domain, then use “Practice this domain” to answer questions from that domain only. Domains with a larger weight appear more often on the exam.

01

Design Solutions for Organizational Complexity

26%

Learning goals

  • Connect many VPCs and on-premises sites with Transit Gateway, Direct Connect, Site-to-Site VPN and PrivateLink, and design hybrid DNS with Route 53 Resolver
  • Design multi-account governance and log aggregation with Organizations, Control Tower, SCPs, RCPs and delegated administrators
  • Design organization-wide access and encryption with IAM Identity Center and external identity providers, cross-account roles and KMS key policies
  • Design DR by RPO/RTO (including Elastic Disaster Recovery) and cost visibility with tags, Cost Categories and purchase options

Key points

  • With dozens of VPCs or more, use a Transit Gateway hub-and-spoke rather than a mesh of peering. Separate route tables enable isolation such as keeping production and development VPCs from talking to each other while both reach a shared services VPC (peering does not route transitively).
  • On-premises connectivity: use Direct Connect when you need stable bandwidth and low latency (redundant with two or more connections at multiple locations), and Site-to-Site VPN when you need it right away or as a backup line. Reach VPCs in multiple Regions through a Direct Connect gateway.
  • Hybrid DNS: to resolve on-premises names from AWS, use Route 53 Resolver outbound endpoints and forwarding rules; to resolve AWS private hosted zones from on-premises, use inbound endpoints. Forwarding rules can be shared with other accounts through RAM.
  • To expose just one service in another VPC or account, use PrivateLink (endpoint services). It works even with overlapping CIDRs and does not connect the whole VPC.
  • Multi-account: separate log archive and audit accounts with a Control Tower landing zone, and aggregate with an organization CloudTrail trail and a Config aggregator. Make a security account the delegated administrator for GuardDuty and Security Hub to manage all accounts centrally.
  • SCPs set the maximum permissions available to IAM principals in accounts, and RCPs set the maximum access to resources (S3, KMS and so on), organization-wide. Neither grants permissions.
  • Cost visibility: allocate costs per business unit with cost allocation tags (values standardized with tag policies) and Cost Categories, and alert on overruns and anomalies with Budgets and Cost Anomaly Detection. With consolidated billing, Savings Plans and RI discounts are shared within the organization.

Workplace uses

  • General AffairsSet up about 40 accounts of 8 group companies with Control Tower and consolidate each company's VPCs on Transit Gateway. Connect to the head-office data center with Direct Connect at two locations.
  • AccountingStandardize “CostCenter” tag values with a tag policy and allocate costs by business unit every month with Cost Categories. Notify department heads at 80% of budget.
  • SalesExpose only the order API to sales agents' accounts through PrivateLink, without connecting them to the entire internal network.

Exam objectives (from the exam guide)

  • 1.1 ネットワーク接続戦略を設計する。
  • 1.2 セキュリティコントロールを規定する。
  • 1.3 信頼性とレジリエンスに優れたアーキテクチャを設計する。
  • 1.4 マルチアカウント AWS 環境を設計する。
  • 1.5 コスト最適化と可視化の戦略を決定する。

Learning resources

02

Design for New Solutions

29%

Learning goals

  • Design rollback-capable deployment strategies (blue/green, canary, rolling) with CloudFormation (StackSets, change sets), CDK, CodePipeline and CodeDeploy
  • Design Multi-AZ and multi-Region data replication (Aurora Global Database, DynamoDB global tables, S3 replication) and DR testing
  • Combine least privilege, WAF, Shield Advanced, Network Firewall, encryption and key management, and patch management from requirements
  • Balance performance and cost goals with purpose-built databases, caching, edge, loosely coupled architectures and purchase options

Key points

  • Deployment: codify infrastructure (CloudFormation, CDK) and use StackSets to deploy the same configuration to multiple accounts and Regions. For production changes, review the diff in a change set before executing.
  • Blue/green runs old and new environments side by side and switches, so you can roll back immediately. Canary sends only part of the traffic to the new version to observe it. RDS Blue/Green Deployments can be used for database engine upgrades.
  • Let managed services handle it: a common question is the decision to prefer RDS, Aurora, Fargate, Lambda and so on over self-managed EC2 to reduce the operational burden of patching and scaling.
  • Business continuity: choose a replication method that meets RPO/RTO. Aurora Global Database for cross-Region relational DBs, DynamoDB global tables for multi-Region NoSQL writes, and S3 replication for objects (S3 RTC for time-bound replication).
  • Attack protection for large web apps: CloudFront + AWS WAF (managed rules, rate-based rules) + Shield Advanced. Use Network Firewall to inspect traffic within VPCs and Gateway Load Balancer for third-party virtual appliances.
  • Reliability: loose coupling (SQS, SNS, EventBridge, Step Functions), retries and idempotency, and raising service quotas in advance. Choose Route 53 routing (latency, geolocation, failover, weighted) by requirements.
  • Performance and cost: choose purpose-built databases (DynamoDB, ElastiCache, OpenSearch, Timestream, Neptune and so on) by use case. Lower costs while maintaining performance with Graviton, Savings Plans, Spot and Auto Scaling predictive scaling.

Workplace uses

  • SalesBuild a new order site for agents nationwide with CloudFront, WAF, ALB and Aurora, and release blue/green from CodePipeline. If problems arise, roll back to the old version within minutes.
  • AccountingIn a new expense system, replicate Aurora in Tokyo to Osaka with Aurora Global Database and run a Region-switch DR test once a quarter.
  • General AffairsAdopt OpenSearch Service to search entry/exit logs and ElastiCache to cache presence status, using databases suited to each purpose.

Exam objectives (from the exam guide)

  • 2.1 ビジネス要件を満たすデプロイ戦略を設計する。
  • 2.2 事業の継続性を確保するソリューションを設計する。
  • 2.3 要件に基づいてセキュリティコントロールを決定する。
  • 2.4 信頼性の要件を満たす戦略を策定する。
  • 2.5 パフォーマンス目標を満たすソリューションを設計する。
  • 2.6 ソリューションの目標と目的を達成するためのコスト最適化戦略を決定する。

Learning resources

03

Continuous Improvement for Existing Solutions

25%

Learning goals

  • Set up monitoring, alerting and automatic remediation with CloudWatch (metrics, logs, Synthetics), EventBridge and Systems Manager Automation
  • Continuously improve security of existing environments with Config rules and remediation actions, Security Hub, Access Analyzer and Secrets Manager
  • Identify bottlenecks and improve performance by revisiting caching, read replicas, edge and instances
  • Remove single points of failure, enable self-healing and fault injection testing (FIS), and drive cost improvement with Compute Optimizer, Trusted Advisor and S3 Storage Lens

Key points

  • Operational improvement is “visibility → alerts → automatic remediation.” Run Systems Manager Automation runbooks from CloudWatch alarms or EventBridge rules to automate recovery with defined procedures.
  • If manual deployments or configuration changes cause outages, propose CI/CD and codification, review with change sets, and moving to rolling or blue/green deployments.
  • Continuous security improvement: detect configuration drift with Config rules and fix it automatically with remediation actions. Aggregate findings in Security Hub, and use Access Analyzer to find externally shared resources and unused permissions.
  • Move hard-coded credentials to Secrets Manager (automatic rotation) or Parameter Store. Replacing SSH bastions to EC2 with Session Manager keeps ports closed and leaves session logs.
  • Performance: first identify the bottleneck (CPU, IOPS, DB connections, network) with metrics, then act. For read-heavy DBs use caching and read replicas; for users worldwide use CloudFront or Global Accelerator.
  • Reliability: eliminate single points of failure such as a single AZ, single instance or single line. Use AWS FIS to deliberately inject faults and verify in a production-like environment that recovery procedures and mechanisms work.
  • Cost: find oversized resources with Compute Optimizer, idle resources with Trusted Advisor and unneeded data with S3 Storage Lens. Also review network costs such as transfers via NAT gateways and inter-AZ traffic.

Workplace uses

  • General AffairsAn intranet portal EC2 instance stops from memory exhaustion once a week. Monitor memory with the CloudWatch agent, and at a threshold Systems Manager Automation automatically restarts it and collects logs.
  • AccountingA database password was written in plain text in the accounting system's configuration. Move it to Secrets Manager with automatic rotation, and detect recurrence with a Config rule.
  • SalesThe quoting system responds slowly. X-Ray shows database calls are the bottleneck, and adding ElastiCache improves response times.

Exam objectives (from the exam guide)

  • 3.1 全体的な運用上の優秀性を高めるための戦略を作成する。
  • 3.2 セキュリティを向上させるための戦略を決定する。
  • 3.3 パフォーマンスを改善するための戦略を決定する。
  • 3.4 信頼性を向上させるための戦略を決定する。
  • 3.5 コスト最適化の機会を特定する。

Learning resources

04

Accelerate Workload Migration and Modernization

20%

Learning goals

  • Assess the portfolio with Migration Hub and Application Discovery Service, and set migration priorities and waves using the 7 Rs (rehost, replatform, refactor and so on) and TCO
  • Choose migration tools for apps, databases and data from Application Migration Service, DMS, SCT, DataSync, Transfer Family and the Snow Family
  • Choose targets such as EC2, Elastic Beanstalk, ECS/EKS/Fargate, EBS, EFS, FSx, S3, Storage Gateway, RDS and DynamoDB
  • Propose modernization to loosely coupled, serverless designs with Lambda, containers, purpose-built databases and SQS/SNS/EventBridge/Step Functions

Key points

  • Migration assessment: collect server configuration, utilization and dependencies with Application Discovery Service (agent-based or agentless) and track progress centrally in Migration Hub. Move strongly dependent servers in the same wave.
  • The 7 Rs: rehost (move as is), replatform (small changes such as replacing the OS or DB with managed services), refactor (rebuild cloud-native), repurchase (move to SaaS), relocate, retain and retire. With a tight deadline, rehost first and optimize later.
  • Migrate servers with Application Migration Service (continuous block-level replication with short cutover). Migrate databases with DMS (full load + CDC while staying online); between different engines, convert schemas with SCT (or DMS Schema Conversion).
  • Data migration: DataSync to continuously sync files over the network, Transfer Family for SFTP/FTPS partner integration, and Snowball Edge for thin lines or hundreds of TB or more. Use Storage Gateway to keep using data from on-premises during migration.
  • Choosing targets: FSx for Windows File Server for Windows file shares, FSx for NetApp ONTAP for migrations from NetApp, and EFS for Linux shares. If commercial DB licenses or special OS access are needed, RDS Custom or self-managed on EC2 are options.
  • Modernization: Lambda and EventBridge for batch and event processing, Step Functions for long-running workflows, and containerizing parts of a monolith onto ECS/Fargate (the strangler pattern).
  • Replacing with purpose-built databases: ElastiCache/DynamoDB for sessions and fast lookups, OpenSearch for full-text search, and Aurora Serverless v2 for relational DBs with unpredictable load.

Workplace uses

  • General AffairsMigrate 120 servers in the 9 months before the data center contract renews. Investigate dependencies with Application Discovery Service and rehost in waves with Application Migration Service.
  • AccountingMigrate the on-premises Oracle accounting DB to Aurora PostgreSQL with DMS full load and CDC. Convert the schema with SCT and cut over avoiding month-end.
  • SalesReplace a nightly batch that built prospect lists with an EventBridge schedule, Lambda and Step Functions, eliminating server operations.

Exam objectives (from the exam guide)

  • 4.1 移行が可能な既存のワークロードとプロセスを選択する。
  • 4.2 既存ワークロードの最適な移行アプローチを決定する。
  • 4.3 既存ワークロードの新しいアーキテクチャを決定する。
  • 4.4 モダナイゼーションと機能強化の機会を決定する。

Learning resources

When you finish reading

Notes (unofficial, original questions, no pass guarantee)