Subscribe
Share
Share
Embed
Ready to earn your AWS Certified Cloud Practitioner credential? Our prepcast is your ultimate guide to mastering the fundamentals of AWS Cloud, including security, cost management, core services, and cloud economics. Whether you're new to IT or looking to expand your cloud knowledge, this series will help you confidently prepare for the exam and take the next step in your career. Produced by BareMetalCyber.com, your trusted resource for expert-driven cybersecurity education.
When approaching AWS pricing, it helps to think in terms of three pillars: pay-as-you-go flexibility, commitment-based discounts, and efficiency improvements through design. AWS’s model is intentionally broad, giving you ways to align spend with business priorities. The trade-off is that you need to understand the levers. On-Demand is flexible but at list price. Commitments with Reserved Instances and Savings Plans lower costs in exchange for steady use. Spot Instances slash prices further for workloads that can tolerate interruptions. Layered on top are architectural choices like caching or lifecycle policies, and governance controls like budgets and tagging. The exam doesn’t test fine-grained dollar amounts, but it does test whether you know which model or tool applies to a given workload profile.
On-Demand pricing is the simplest model. You pay only for what you use, with no long-term commitments. It’s ideal for unpredictable workloads or for starting new projects when demand is unknown. The trade-off is that you pay full list price per hour or per second, depending on the service. For example, if you launch an m5.large instance for a few hours and terminate it, you only pay for those hours. On-Demand provides agility, but if you leave resources running 24×7, the bill quickly grows. Exam questions often use phrases like “unknown usage pattern” or “new service” to signal On-Demand as the correct answer.
Reserved Instances, or RIs, provide discounts in exchange for committing to one- or three-year terms. Standard RIs give the deepest savings but are tied to a specific instance family, Region, and operating system. Convertible RIs are more flexible—you can exchange them for other families—but provide smaller discounts. RIs are most appropriate for steady workloads where usage is predictable, such as a company running the same EC2 fleet year-round. The exam cue is “steady baseline” or “predictable usage,” which always maps to RIs or Savings Plans.
Savings Plans evolved from RIs to provide broader flexibility. Compute Savings Plans apply to any compute usage—EC2, Fargate, or Lambda—as long as you commit to a consistent dollar spend per hour. EC2 Instance Savings Plans are narrower, applying to a specific instance family in a Region, similar to RIs. Both provide discounts for one- or three-year commitments, with Compute Savings Plans offering the most flexibility. For example, if your workload might shift between EC2 and Fargate, a Compute Savings Plan is the safer choice. The exam often contrasts RIs and Savings Plans, so remember: RIs are specific, Savings Plans are flexible.
Spot Instances represent another model entirely: AWS sells spare capacity at deep discounts, sometimes 90 percent off On-Demand. The catch is that AWS can reclaim the capacity with short notice. Spot is ideal for interruption-tolerant workloads like rendering, data crunching, or test environments. The exam cue is “fault-tolerant batch” or “checkpoint progress,” which always points to Spot. Never pick Spot if the requirement says “mission-critical” or “no downtime.”
The Free Tier is AWS’s entry point for new users. It includes a 12-month free allocation for certain resources, such as 750 hours of t2.micro or t3.micro EC2 usage, plus “always free” services like a million Lambda requests per month. Some services also have trial-based free tiers. Exam questions may ask about the Free Tier, but the key is to know that it has limits—it’s not unlimited production use. “Learn, experiment, and test” is the right framing.
Data transfer pricing is another cost driver. Inbound data to AWS is free, but outbound traffic—to the internet or across Regions—incurs charges. Even inter-AZ transfers in the same Region may cost money. To reduce transfer costs, services like CloudFront, caching, and PrivateLink are common. The exam might frame this as “runaway data egress costs,” with CloudFront as the correct answer.
Storage pricing follows a tiered model. In S3, Standard is for frequent access, Standard-IA and One Zone-IA for infrequent use, Glacier and Glacier Deep Archive for archival, and Intelligent-Tiering for unpredictable access. Lifecycle policies move data automatically between these classes. For block storage, EBS pricing depends on type: gp3 for balanced performance, io2 for high IOPS, st1 for throughput, and sc1 for low-cost archival. The exam rarely asks about prices but will test whether you recognize which class or type matches the usage pattern.
Compute cost drivers are straightforward: instance family, instance size, Region, and hours used. Larger instances cost more, Regions differ in price, and running instances continuously multiplies spend. For example, the same m5.large costs less in US East than in Asia Pacific. The exam expects you to know that “type, size, Region, and hours” are the main variables—not obscure formulas.
Serverless services like Lambda follow a different pricing model: you pay for requests and duration. For Lambda, that means number of invocations and how long the function runs, measured in milliseconds of execution time. DynamoDB charges for provisioned or on-demand read/write capacity units. These models emphasize efficiency—you pay when something runs, not for idle time. Exam cues like “millisecond execution” or “pay per request” map directly to serverless services.
Managed services follow predictable pricing patterns: RDS charges for instance hours, storage, and backups. DynamoDB charges for read/write units and storage. ElastiCache charges for node class and hours. The key takeaway is that managed services price by resources consumed, not by hidden licenses. For the exam, if the scenario asks about “removing database patching and backups,” the right answer is RDS or Aurora, and pricing follows the same commitment/on-demand patterns as compute.
Support plans also fall into the pricing conversation. Basic support is free, covering documentation and forums. Developer support provides guidance for non-production environments. Business support, a common enterprise choice, includes 24×7 access to engineers and faster response times. Enterprise support adds a Technical Account Manager and proactive reviews. On the exam, “mission-critical with 24×7 support” points to Business or Enterprise, while “experimentation” points to Basic or Developer.
The AWS Pricing Calculator is an important tool for estimating costs before deploying workloads. It allows you to configure resources and see projected charges, helping with forecasting. The exam won’t ask you to use it, but it may ask which tool helps you estimate costs before building—always the Pricing Calculator.
Finally, AWS cost visibility tools include Budgets, Cost Explorer, and the Cost and Usage Report (CUR). Budgets set thresholds and trigger alerts. Cost Explorer shows trends and allocation by tag or service. CUR provides the most detailed, line-item charge data, often used with Athena for analysis. On the exam, “alerts” means Budgets, “trends” means Cost Explorer, and “granular detail” means CUR.
For more cyber related content and books, please check out cyber author dot me. Also, there are other prepcasts on Cybersecurity and more at Bare Metal Cyber dot com.
When matching pricing models to workloads, predictability is the key decision point. If the workload runs steadily 24×7 with minimal change, Reserved Instances or Savings Plans provide the best value. If the workload is unpredictable or new, On-Demand is the safest starting point until patterns emerge. For spiky, fault-tolerant jobs, Spot Instances offer unmatched discounts. The exam often phrases this as “predictable versus unpredictable usage” or “interruption-tolerant versus critical.” Your answer should always align commitment with stability and flexibility with uncertainty.
Baselines and bursts are another common exam distinction. Baseline capacity should be covered by Reserved Instances or Savings Plans, ensuring savings on the steady portion of usage. Bursty or overflow capacity can run On-Demand or Spot, giving elasticity without waste. For example, a retailer might cover its steady ecommerce load with Savings Plans while using On-Demand for holiday peaks. The exam cues are “steady workload” for commitment models and “variable spikes” for On-Demand or Spot.
Spot Instances require an interruption tolerance checklist. If the scenario mentions batch jobs, rendering, analytics, or any task with checkpoints and retries, Spot is correct. If it says “mission-critical,” “database,” or “must not be interrupted,” Spot is the wrong choice. The exam will test whether you know Spot’s deep savings come with trade-offs. Always map “interruption-tolerant” to Spot, and “must always be available” to Reserved, Savings Plans, or On-Demand.
Choosing between Savings Plans and Reserved Instances comes down to flexibility versus specificity. RIs lock savings to a family, Region, and OS, giving the highest discounts. Savings Plans, especially Compute Savings Plans, apply more broadly across EC2, Fargate, and Lambda. If the workload might change services or families, Savings Plans are better. If the workload is fixed for years, RIs maximize discounts. The exam often contrasts these, so remember: RIs are narrow but deep, Savings Plans are wide and flexible.
Capacity reservations also appear on the exam. They guarantee capacity in a given AZ, ensuring instances can always launch. They are separate from pricing commitments but can be combined with RIs or Savings Plans. If the scenario says “must guarantee instance availability” rather than “reduce cost,” the answer is a capacity reservation. The exam sometimes blends these concepts, so be careful to distinguish guaranteed capacity from pricing discounts.
Data transfer cost reduction comes from smarter architecture. CloudFront caches content at the edge, reducing repeated egress. PrivateLink and VPC endpoints keep traffic inside AWS, avoiding internet routes. Exam questions often describe “unexpectedly high data transfer costs,” with CloudFront or PrivateLink as the correct answers. The key pattern: reduce egress by keeping data close to users or inside the AWS network.
Storage optimization is automated with lifecycle policies. They transition data into cheaper classes like IA or Glacier without manual work. If the scenario says “data must be retained but rarely accessed,” lifecycle rules are the answer. This is a favorite exam topic because it ties directly to cost efficiency with minimal effort.
Rightsizing and scheduling address idle compute costs. Exam cues like “low utilization” or “idle resources at night” point to shutting down or downsizing instances. For example, a dev environment can be scheduled to stop after hours. These practices ensure you pay only for what you need, a recurring cost optimization theme.
Governance is enforced through budgets, alerts, and tagging. Budgets set thresholds and provide proactive alerts. Tagging ensures costs can be attributed to teams or projects. Without governance, bills become opaque. The exam may phrase this as “allocate costs to departments” or “get alerted before overages,” which point to tags and budgets respectively.
Forecasting and variance analysis ensure cost predictability. Budgets and Cost Explorer both support forecasting, while the Cost and Usage Report provides detailed variance analysis. Exam cues like “drill down to resource-level detail” point to CUR, while “analyze historical trends” means Cost Explorer.
Marketplace and licensing considerations appear lightly on the exam. AWS Marketplace lets you buy software and licenses billed through AWS. Some enterprise scenarios highlight BYOL (bring your own license) models. The exam doesn’t go deep here, but you should know Marketplace charges show up on consolidated bills.
Cross-account consolidated billing is a powerful cost lever. It rolls multiple accounts into one payer, pooling discounts from Reserved Instances or Savings Plans. Exam questions may describe “multiple accounts with shared discounts,” and the answer is always consolidated billing through AWS Organizations.
The exam lens on pricing is straightforward: match the scenario to the simplest pricing lever. If the requirement is steady baseline, choose RI or Savings Plans. If it’s unpredictable, choose On-Demand. If it’s tolerant of interruptions, Spot is correct. If it’s about forecasting, pick the Pricing Calculator. If it’s about alerts, pick Budgets. The key is to keep answers simple—don’t overbuild or pick tools that go beyond the question’s scope.
In conclusion, AWS pricing models balance flexibility, commitment, and efficiency. Commit where workloads are predictable, use flexible models for bursts, and apply architectural practices like caching and lifecycle policies to minimize unnecessary cost. Governance through budgets, tagging, and forecasting ensures accountability. For the exam, remember: steady equals Reserved/Savings Plans, unpredictable equals On-Demand, tolerant equals Spot. Combined with governance, these patterns deliver predictable spend in both study and practice.