AWS Certified Solutions Architect Design Cost-Optimized Architectures Syllabus

Design Cost-Optimized Architectures syllabus — full chapter and topic list

Expand any chapter to see its topics and sub-topics. This is the whole examinable outline for Design Cost-Optimized Architectures in AWS Certified Solutions Architect, not a summary of it.

1. 1

   Cost-Effective Compute

   3 topics

   • EC2 Pricing Models
     • On-Demand, Reserved Instances, and Savings Plans
     • Spot Instances and interruption handling
     • Right-sizing and instance scheduling
   • Serverless Cost Optimization
     • Pay-per-use Lambda and Fargate economics
     • Eliminating idle capacity with event-driven design
   • Cost Visibility for Compute
     • Compute Optimizer recommendations
     • Tagging strategy for cost allocation
2. 2

   Cost-Effective Storage and Data Transfer

   3 topics

   • Storage Tiering for Cost
     • S3 storage class selection and Glacier tiers
     • Intelligent-Tiering for unpredictable access
     • Lifecycle transitions and expiration
   • Data Transfer Cost Management
     • Inter-AZ, inter-Region, and egress charges
     • Using CloudFront and VPC endpoints to cut transfer cost
   • Database Cost Optimization
     • Aurora Serverless v2 scaling to demand
     • DynamoDB on-demand vs provisioned capacity
3. 3

   Cost Governance and Monitoring

   3 topics

   • Cost Management Tooling
     • AWS Cost Explorer analysis
     • AWS Budgets and budget actions
     • Cost and Usage Report (CUR)
   • Optimization Recommendations
     • AWS Trusted Advisor cost checks
     • Identifying and removing unused resources
   • Organizational Cost Controls
     • Consolidated billing and volume discounts
     • Reserved Instance and Savings Plan sharing across accounts

Design Cost-Optimized Architectures flashcards for AWS Certified Solutions Architect

24 of 51 cards from the Design Cost-Optimized Architectures deck — real questions with worked answers.

1. What are the four main EC2 purchasing (pricing) options?

   On-Demand, Reserved Instances (RIs), Savings Plans, and Spot Instances (plus Dedicated Hosts/Instances for compliance).

2. When should you choose On-Demand EC2 pricing?

   For short-term, spiky, or unpredictable workloads that can't be interrupted, where you want no upfront commitment and pay per second/hour.

3. What discount can Spot Instances provide versus On-Demand, and what is the trade-off?

   Up to ~90% off On-Demand, but AWS can reclaim the instance with a 2-minute warning when capacity is needed — best for fault-tolerant, stateless, flexible workloads.

4. What is the maximum commitment term for Reserved Instances and Savings Plans?

   1 year or 3 years; longer terms and more upfront payment give bigger discounts.

5. What are the three RI/Savings Plan payment options?

   All Upfront (largest discount), Partial Upfront, and No Upfront (smallest discount).

6. Difference between Standard and Convertible Reserved Instances?

   Standard RIs give a higher discount but limited changes; Convertible RIs give a lower discount but can be exchanged for RIs with different instance family, OS, or tenancy.

7. Compare Compute Savings Plans vs EC2 Instance Savings Plans.

   Compute Savings Plans (up to ~66% off) apply across any region, instance family, OS, tenancy, and even Fargate/Lambda; EC2 Instance Savings Plans (up to ~72% off) are locked to a specific instance family in a region but allow size/OS/AZ changes.

8. What is a Spot Fleet / EC2 Fleet, and why use it for cost optimization?

   It launches a target capacity across multiple instance types and AZs mixing Spot and On-Demand, improving availability and getting the lowest-cost capacity automatically.

9. How does Lambda pricing work?

   You pay per request plus compute duration measured in GB-seconds (allocated memory x execution time, billed per millisecond); no charge when idle.

10. In Lambda, how does increasing memory affect cost and performance?

    More memory also proportionally increases CPU/network, so functions often run faster — the right-sized memory can lower total GB-seconds and cost despite a higher per-ms rate.

11. What is AWS Lambda Power Tuning used for?

    A tool/state machine that runs a Lambda at different memory settings to find the configuration with the best cost-performance balance.

12. How can Provisioned Concurrency affect Lambda cost?

    It keeps functions warm to avoid cold starts but adds a fixed hourly charge for kept-warm capacity, so it only saves money/improves UX for predictable steady traffic.

13. Why is Fargate often cheaper than EC2 for spiky container workloads?

    Fargate charges only for the vCPU and memory provisioned while tasks run (no idle EC2 hosts to manage), so you avoid paying for unused cluster capacity.

14. What is Compute Optimizer and how does it help cost optimization?

    It uses ML on CloudWatch metrics to recommend optimal EC2 instance types/sizes, Auto Scaling group configs, EBS volumes, and Lambda memory to right-size and cut cost.

15. Which CloudWatch metric is NOT collected by default and matters for right-sizing EC2?

    Memory utilization — it requires the CloudWatch agent; without it, right-sizing based only on CPU can be misleading.

16. What tool gives the most granular view of where compute (and all) spend goes, and what dimensions can it group by?

    AWS Cost Explorer — group by service, region, instance type, tag, account, usage type, etc., over time.

17. How do cost allocation tags improve compute cost visibility?

    User-defined or AWS-generated tags (once activated in Billing) let you break down and chargeback costs by project, team, environment, or owner in Cost Explorer and reports.

18. List the S3 storage classes from highest to lowest cost (frequent to archive).

    S3 Standard > Standard-IA / One Zone-IA > Glacier Instant Retrieval > Glacier Flexible Retrieval > Glacier Deep Archive (Intelligent-Tiering auto-moves between them).

19. What does S3 Intelligent-Tiering do and what is its trade-off?

    Automatically moves objects between access tiers based on usage to cut cost with no retrieval fees; you pay a small per-object monitoring/automation charge — ideal for unknown/changing access patterns.

20. Key cost/durability difference between S3 Standard-IA and One Zone-IA?

    Both have lower storage cost than Standard plus a retrieval fee; One Zone-IA is ~20% cheaper but stores data in a single AZ, so it lacks multi-AZ resilience.

21. What is S3 Glacier Deep Archive best for, and what is its retrieval time?

    The lowest-cost class for long-term archival (e.g., 7-10 year retention) rarely accessed; standard retrieval takes around 12 hours.

22. How do S3 Lifecycle policies reduce cost?

    They automatically transition objects to cheaper classes after a set number of days and/or expire (delete) old objects and incomplete multipart uploads.

23. Which EBS volume type is cheapest, and for what workloads?

    Cold HDD (sc1) is the lowest cost, for infrequently accessed, throughput-oriented data; st1 (Throughput HDD) suits frequent big sequential workloads like logs/big data.

24. How can switching from gp2 to gp3 EBS save money?

    gp3 is ~20% cheaper per GB than gp2 and lets you provision IOPS and throughput independently of size, so you don't overpay on capacity just to get performance.

See more Design Cost-Optimized Architectures flashcards →

Planning Design Cost-Optimized Architectures for AWS Certified Solutions Architect

Design Cost-Optimized Architectures (AWS Certified Solutions Architect) FAQ