4.4 Additional Billing Concepts and Services
Key Takeaways
- AWS Pricing Calculator estimates costs for AWS architectures before you deploy — useful for budgeting and proposals.
- AWS Compute Optimizer analyzes resource utilization and recommends optimal EC2 instances, EBS volumes, and Lambda configurations.
- AWS Marketplace is a digital catalog of third-party and AWS software that can be purchased, deployed, and managed from one place.
- The Total Cost of Ownership (TCO) calculation should include ALL on-premises costs: hardware, software, facilities, staff, and overprovisioning.
- AWS offers volume-based discounts — the more you use certain services (like S3), the less you pay per unit.
Quick Answer: Beyond the core tools, the CLF-C02 expects you to know the AWS Pricing Calculator (estimate before deploying), Compute Optimizer (ML right-sizing), AWS Marketplace (buy third-party software), volume discounts (cheaper per unit as usage grows), and per-service billing quirks for EC2, S3, Lambda, and RDS.
AWS Pricing Calculator
The AWS Pricing Calculator (calculator.aws) builds cost estimates for AWS use cases before you deploy anything.
- Estimate monthly cost for individual services or whole architectures
- Compare pricing across Regions
- Model On-Demand vs. Reserved vs. Savings Plans scenarios
- Generate shareable estimates for proposals and budgets, and export to CSV
On the Exam: Pricing Calculator = estimate BEFORE deployment. Cost Explorer = analyze AFTER deployment. This before/after split is the single most-tested billing distinction.
AWS Compute Optimizer
AWS Compute Optimizer uses machine learning on real utilization data to recommend optimal resources.
| Resource | Recommendation |
|---|---|
| EC2 instances | Right-size up or down from CPU, memory, network |
| EC2 Auto Scaling groups | Optimal instance types and sizes |
| EBS volumes | Right-size volume type and size |
| Lambda functions | Optimal memory allocation |
| ECS on Fargate | Right-size CPU and memory |
Compared with Trusted Advisor's broad checks, Compute Optimizer gives deeper, ML-based right-sizing analysis specifically for compute.
AWS Marketplace
AWS Marketplace is a digital catalog to find, buy, and deploy software from thousands of independent vendors.
What you can buy: pre-configured AMIs, SaaS subscriptions, data products, professional services, and container images. Pricing models include pay-as-you-go, annual subscriptions, free trials, and bring-your-own-license (BYOL). Purchases appear on your consolidated AWS bill, simplifying procurement and vendor management.
Volume-Based Discounts
Many services use tiered pricing — the more you use, the lower the per-unit rate.
Example: S3 storage tiers (illustrative)
| Tier | Per GB-Month |
|---|---|
| First 50 TB | Higher rate |
| Next 450 TB | Lower rate |
| Over 500 TB | Lowest rate |
The same tiered model applies to data transfer and other services. Combined with consolidated billing, usage across all accounts in an organization aggregates to reach the lower tiers sooner.
Service-Specific Pricing Notes
EC2 Pricing Reminders
- Per-second billing for Linux instances (60-second minimum); plain Windows historically bills per hour
- You are charged while an instance is running, not when it is stopped
- EBS volumes keep accruing charges even when the instance is stopped (persistent storage)
- Elastic IPs are free while associated with a running instance, but charged when idle/unassociated
S3 Pricing Reminders
- Storage per GB-month; requests per 1,000 (PUT/GET/LIST); outbound data transfer per GB
- Lifecycle transitions carry a small per-object cost; inbound transfer is free
Lambda Pricing
- Charged per 1M requests plus GB-seconds of duration
- Always-Free tier: 1M requests + 400,000 GB-seconds per month
RDS Pricing
- Instance hours by size, database storage per GB-month, backup storage beyond the provisioned size, and standard data-transfer rules
Total Cost of Ownership (TCO)
When comparing cloud to on-premises, a complete TCO analysis must include the hidden on-premises costs people forget: hardware refresh, software licenses, power and cooling, data-center real estate, staffing, disaster-recovery sites, and overprovisioning for peak demand. Many of these shrink or disappear in the cloud's pay-as-you-go model, which is why TCO comparisons usually favor cloud once all costs are counted.
Cost Optimization Best Practices
- Right-size with Compute Optimizer recommendations
- Use Savings Plans / Reserved Instances for predictable workloads
- Use Spot Instances for fault-tolerant workloads
- Delete waste — unattached EBS volumes, idle Elastic IPs, old snapshots
- Use S3 lifecycle policies to tier cold data automatically
- Use Auto Scaling to scale down off-peak
- Monitor with Cost Explorer to catch trends early
- Set Budgets and alerts to stop overruns before they grow
- Use consolidated billing to aggregate for volume discounts
- Tag everything to attribute spend to teams and projects
On the Exam: "Identify idle/unused resources" most often points to Trusted Advisor (cost-optimization checks); "recommend a better instance size" points to Compute Optimizer; "estimate before building" points to the Pricing Calculator.
Putting the Cost Tools in Order
A useful way to remember which tool to reach for is to sequence them across a workload's life. Plan: estimate with the Pricing Calculator and prototype on the Free Tier. Commit: once usage is predictable, buy Savings Plans or Reserved Instances and run interruptible work on Spot. Operate: tag everything, set Budgets with alerts, and let anomaly detection flag spikes. Review: open Cost Explorer for trends, accept Compute Optimizer right-sizing, and pull the CUR for resource-level detail. Govern: use consolidated billing to aggregate discounts across accounts.
Most "most cost-effective" exam questions are really asking which lever fits the described situation — a steady server (commitment), a batch job that can restart (Spot), cold data (a cheaper S3 class or lifecycle policy), or an oversized instance (right-sizing). Reading the scenario for predictable vs. interruptible workload behavior and hot vs. cold data access usually points straight to the intended answer, and avoids the trap of picking an impressive-sounding but mismatched option.
Finally, remember that the cheapest architecture is often the one that simply turns off what it does not need: deleting unattached volumes, releasing idle Elastic IPs, and scaling non-production environments down outside business hours frequently saves more than any pricing model alone.
Which tool should a company use to estimate the cost of an AWS architecture BEFORE deploying it?
Which statement about Amazon EC2 billing is correct?
How does AWS Marketplace benefit customers?
A company wants to identify if their EC2 instances are over-provisioned and recommend the optimal instance type. Which service should they use?
A company wants to stop paying for unused Elastic IPs and unattached EBS volumes. Which AWS service can identify these wasted resources?
Which of the following is a benefit of consolidated billing in AWS Organizations?
A team member accidentally launches a very expensive EC2 instance type. What tool could have prevented the surprise cost?
Which statement about the AWS Free Tier is correct?