Pair your devices with a code and playback position follows you: pause on this device, hit resume on the other. Position is saved to the site every minute and on pause.
Open this panel on your other device and enter the same code.
Domain 4 is 20% of scored content — the smallest domain, and the one people most often leave for last. That's a mistake for a simple reason: cost is the tie-breaker in the other three. When two answers both meet a security, resilience or performance requirement, the exam's "most cost-effective" wording picks between them. So Domain 4's facts decide questions that are nominally about Domains 1–3.
The domain has four task statements, one per resource type:
And one thing they share: the first two "Knowledge of" bullets under all four task statements are identical — cost allocation tags, multi-account billing, Cost Explorer, Budgets, and the Cost and Usage Report. So lesson 1 covers the tools once, and the other four lessons use them.
The cheapest answer is the one that still meets the requirement — not the cheapest thing on the page.
WHAT THE STEM SAYS THE LEVER THE TRAP
────────────────── ───────── ────────
steady, all year → commit (Savings Plans / RIs) → committing to what will change
interruptible, flexible → Spot (2-minute notice) → Spot for a stateful singleton
unknown / spiky → on-demand / serverless → provisioning for the peak
rarely read → colder storage class → 30/90/180-day and 128 KB minimums
bytes through a NAT → gateway endpoint (no charge) → paying NAT per GB for S3 traffic
bytes out to the world → CloudFront → serving egress from the origin
⚠️ Every cheap option has a condition attached. Spot needs interruptibility. Glacier needs you to wait. One Zone-IA needs the data to be reproducible. A single NAT gateway needs you to accept losing an AZ's outbound path. The exam's distractors are the cheap options whose condition the stem quietly violates.
| # | Lesson | Task statement | Read | Listen |
|---|---|---|---|---|
| 1 | The cost management toolkit — see it, allocate it, cap it, fix it | all | 28 min | 10 min |
| 2 | Storage cost — classes, minimums, lifecycle, and who pays | 4.1 | 32 min | 10 min |
| 3 | Compute cost — commit, interrupt, right-size, or go serverless | 4.2 | 32 min | 10 min |
| 4 | Database cost — engine, capacity mode, retention, and the cache in front | 4.3 | 28 min | 10 min |
| 5 | Network cost — which bytes you pay for, and the routes that avoid them | 4.4 | 30 min | 9 min |
| Then: Cheat sheet · Lab · Quiz · Interview |
From the SAA-C03 exam guide — Content Domain 4 (verified 2026-09-25), Design Cost-Optimized Architectures — 20% of scored content:
| Task statement | This module |
|---|---|
| 4.1 Design cost-optimized storage solutions | lessons 1, 2 |
| 4.2 Design cost-optimized compute solutions | lessons 1, 3 |
| 4.3 Design cost-optimized database solutions | lessons 1, 4 |
| 4.4 Design cost-optimized network architectures | lessons 1, 5 |
Every "Knowledge of" and "Skills in" item on the Domain 4 page, and where it's taught. Items repeated across task statements are listed once.
| Item (verbatim, abbreviated) | Lesson |
|---|---|
| Cost management service features (cost allocation tags, multi-account billing) — all four | 1 |
| Cost management tools (Cost Explorer, Budgets, Cost and Usage Report) — all four | 1 |
| 4.1 Access options (Requester Pays) | 2 |
| 4.1 Storage services with use cases (FSx, EFS, S3, EBS) | 2 (types in SAA3 1) |
| 4.1 Backup strategies · Selecting backup and/or archival solution | 2, 4 |
| 4.1 Block storage options (HDD, SSD volume types) | 2 |
| 4.1 Data lifecycles · Managing S3 object lifecycles · Selecting the correct data lifecycle | 2 |
| 4.1 Hybrid storage options (DataSync, Transfer Family, Storage Gateway) · Service for data migration | 2 |
| 4.1 Storage access patterns · Storage tiering · Selecting the appropriate storage tier | 2 |
| 4.1 Storage types (object, file, block) · Most cost-effective storage service | 2 (types in SAA3 1) |
| 4.1 Batch uploads vs individual uploads | 2 |
| 4.1 Correct storage size · When storage auto scaling is required | 2 |
| 4.1 Lowest cost method of transferring data to AWS storage | 2, 5 |
| 4.2 Global infrastructure (AZs, Regions) | 3 ⚠️ (Region price variation only) |
| 4.2 Purchasing options (Spot, RIs, Savings Plans) | 3 |
| 4.2 Distributed compute strategies (edge processing) | 3 ⚠️ flagged, not taught |
| 4.2 Hybrid compute options (Outposts) | 3 |
| 4.2 Instance types, families, sizes, virtualization · Selecting family · Selecting size | 3 (families in SAA3 2) |
| 4.2 Optimization of compute utilization (containers, serverless, microservices) | 3 |
| 4.2 Scaling strategies (auto scaling, hibernation) · Scaling methods (horizontal/vertical, hibernation) | 3 (mechanics in SAA2 2) |
| 4.2 Load balancing strategy (ALB / NLB / GWLB) | 3 ⚠️ function in SAA2 3; no ELB pricing fetched |
| 4.2 Cost-effective compute services (Lambda, EC2, Fargate) | 3 |
| 4.2 Required availability for production vs non-production | 3, 5 |
| 4.3 Caching strategies | 4 |
| 4.3 Data retention policies · Backup and retention policies (snapshot frequency) | 4 |
| 4.3 Database capacity planning (capacity units) | 4 |
| 4.3 Database connections and proxies | 4 → SAA2 5 (RDS Proxy) |
| 4.3 Database engines (heterogeneous/homogeneous migrations) · Migrating schemas and data | 4 |
| 4.3 Database replication (read replicas) | 4 → SAA2 5, SAA3 3 |
| 4.3 Database types and services (relational vs non-relational, Aurora, DynamoDB) · DynamoDB vs RDS, serverless | 4 |
| 4.3 Appropriate engine (MySQL vs PostgreSQL) | 4 ⚠️ no cost comparison page found |
| 4.3 Database types (time series, columnar) | 4 |
| 4.4 Load balancing concepts (ALB) | 5 → SAA2 3 |
| 4.4 NAT gateways (NAT instance vs NAT gateway costs) · NAT gateway types (shared vs per AZ) | 5 |
| 4.4 Network connectivity (private lines, dedicated lines, VPNs) · DX vs VPN vs internet | 5 |
| 4.4 Routing, topology, peering (Transit Gateway, VPC peering) | 5 |
| 4.4 Network services (DNS) | 5 ⚠️ no Route 53 pricing fetched |
| 4.4 Routes to minimize transfer cost (Region, AZ, private/public, Global Accelerator, VPC endpoints) | 5 (Global Accelerator ⚠️) |
| 4.4 CDNs and edge caching | 5 |
| 4.4 Reviewing existing workloads for network optimizations | 1, 5 |
| 4.4 Throttling strategy | 5 |
| 4.4 Bandwidth allocation (single vs multiple VPNs, Direct Connect speed) | 5 |
This module cites the HTML exam guide at
docs.aws.amazon.com/aws-certification/latest/solutions-architect-associate-03/, not the
d1.awsstatic.com PDF. Checked again on 2026-09-25: task 4.2's hybrid compute item reads "Hybrid
compute options (for example, AWS Outposts)" — Snowball Edge is not on the HTML page, so this
module doesn't teach it. See ../PLAN.md for the full HTML-vs-PDF diff.
Prices change and vary by Region. Where a lesson quotes a price, it names the Region the pricing page was showing and the fetch date (2026-09-25). Where a lesson can avoid a price, it states the structure instead — "gateway endpoints have no additional charge", "Spot is not covered by Savings Plans" — because structure is what the exam tests and it doesn't drift.
Places where a page I fetched didn't contain something examinable are flagged inline with ⚠️ and the page to read. The largest gaps: internet-egress per-GB tiers and S3 per-request rates (both tables are filled in client-side), an EFS IA minimum duration, Trusted Advisor plan entitlements, hibernation prerequisites, and cost-specific pages for Global Accelerator, ELB and Route 53.
Facts verified 2026-09-25 against the pages cited in each lesson.
Keeps playing into the following modules — 81 min from here to the end of certification prep.