EC2 and pricing spot on demand savings plan reserved instance dedicated host Capacity Reservation

In this lesson, we dive into AWS EC2 pricing and explore six cost models designed to optimize your cloud spend. By understanding how each option works, you can align your infrastructure costs with your application’s growth and usage patterns.

Cloud Growth Phases

Every organization’s cloud journey typically moves through three phases:

Start-up Phase
Experimentation with limited users, resources created and terminated on demand, pay-per-use billing.
Steady Phase
Achieved product-market fit, predictable 24/7 usage, consistent resource requirements.
Growth Phase
Fluctuating hourly demand, advanced analytics and batch processing, high user volume.

The image illustrates the phases of cloud growth: Start-up Phase, Steady Phase, and Growth Phase, each with specific characteristics like on-demand services, regular traffic, and complex functionalities.

As your application evolves, selecting the right EC2 pricing model is key to achieving cost efficiency without sacrificing performance.

Why EC2 Cost Management Matters

Amazon EC2 often represents a significant portion of your AWS bill. Gaining clarity on EC2’s billing mechanisms empowers you to:

Identify idle or underutilized resources
Apply the most cost-effective pricing option
Forecast and control cloud expenditure

The image is about "EC2 Cost Management" and features a pie chart highlighting EC2, along with icons representing AWS Cloud and cost-saving ideas.

AWS provides six pricing models for EC2 instances:

On-Demand Instances
Spot Instances
Savings Plans
Reserved Instances
Dedicated Hosts
On-Demand Capacity Reservations

The image outlines the AWS EC2 pricing model, listing six options: On-Demand Price, Spot, Savings Plans, Reserved Instances, Dedicated Hosts, and On-Demand Capacity Reservation.

To compare these options at a glance, refer to the table below:

Pricing Model	Ideal Use Case	Commitment	Potential Savings
On-Demand	Spiky workloads, dev/test	None	None
Spot Instances	Batch jobs, fault-tolerant apps	None	Up to 90%
Savings Plans	Consistent compute spend	1–3 years	Up to 66% (Compute) / 72% (EC2)
Reserved Instances	Predictable, long-term workloads	1–3 years	Up to 72%
Dedicated Hosts	BYOL, compliance, regulatory requirements	None	Additional RI/Savings Plan discounts
On-Demand Capacity Reservations	Business-critical & high-availability needs	On-demand	None

On-Demand Instances

On-Demand Instances let you pay by the second with no long-term commitments. AWS allocates capacity immediately upon launch, and you’re charged only for the time the instance runs.

Features:

No upfront fees
Per-second billing
Full flexibility to launch and terminate

Use cases:

Development and testing environments
Spiky or unpredictable traffic patterns
Applications with short-term or intermittent workloads

Spot Instances

Spot Instances provide access to spare EC2 capacity at discounts of up to 90% off On-Demand prices. AWS may reclaim Spot Instances with a two-minute warning when capacity is needed elsewhere.

The image illustrates the concept of AWS Spot Instances, showing a diagram of cloud resources with price matching and on-demand requests, alongside icons representing features and use cases.

Key points:

Variable pricing based on supply and demand
Two-minute interruption notice
No upfront commitment

Use cases:

Stateless, fault-tolerant workloads
Batch processing and big data analytics
Flexible start and end times

Note

Spot Instances are ideal for workloads that can gracefully handle interruptions. Consider integrating checkpointing or auto-scaling to reduce impact.

Savings Plans

Savings Plans offer significant discounts in exchange for a constant hourly spend commitment over a one- or three-year term.

The image illustrates AWS Saving Plans, showing a comparison between on-demand and discounted prices, and highlighting Compute and EC2 Instance Savings Plans.

Two plan types:

Compute Savings Plan
- Up to 66% savings
- Applies to any EC2 instance, regardless of family, region, OS, or tenancy
EC2 Instance Savings Plan
- Up to 72% savings
- Locked to a specific instance family in one region
- Flexibility to change instance size, OS, and tenancy within that family

Note

Savings Plans automatically apply your commitment to the lowest-cost usage across eligible instance types.

Reserved Instances

Reserved Instances (RIs) are an older commitment model offering deep discounts in exchange for a one- or three-year term.

The image is an infographic about AWS Reserved Instances, showing a comparison between Standard and Convertible Reserved Instances with savings percentages and flexibility options.

Types of RIs:

Standard Reserved Instances
- Up to 72% off On-Demand pricing
- Modify availability zone, instance size, and network platform
Convertible Reserved Instances
- Up to 66% off On-Demand pricing
- Exchange one RI for another of equal or greater value
- Change instance families, OS, and tenancy

Dedicated Hosts

Dedicated Hosts allocate a physical server solely for your use, enabling you to bring existing software licenses and meet strict compliance requirements.

The image is an infographic about AWS Cloud dedicated hosts, highlighting features like BYOL and reserved instance savings plans, and detailing specifications for the c7g instance family.

Key features:

Full control over sockets and cores
Ideal for BYOL, compliance, and regulatory mandates
Combine with RIs and Savings Plans for extra discounts

Example c7g host capacity:

192 medium instances
96 large instances
48 xlarge instances
24 2xlarge instances

On-Demand Capacity Reservations

On-Demand Capacity Reservations let you reserve compute capacity in a specific Availability Zone, ensuring capacity is available when you need it.

The image is an illustration of AWS On-Demand Capacity Reservation, highlighting features such as capacity assurance, regulatory HA requirements, disaster recovery, and support for RI and savings plans.

Use cases:

Business-critical applications requiring guaranteed capacity
Regulatory or compliance mandates for high availability
Disaster recovery and failover scenarios

Warning

You are billed for your reserved capacity whether or not you run instances. Apply RIs or Savings Plans to offset reservation costs.

With this overview of AWS EC2 pricing models, you can match your workloads to the most cost-effective option and optimize your cloud investment.

Links and References

Watch Video

Watch video content