As an AWS (Amazon Web Services) customer, you’re likely always looking to optimize your cloud spending. While on-demand pricing offers flexibility and control over prices as you pay for what you provision, it’s the most expensive option and can quickly escalate your AWS bill.
To provide more flexibility and cost savings, AWS introduced AWS Savings Plans in 2019. There are three types of Savings Plans: Compute Savings Plans, EC2 Savings Plans, and the recently introduced Amazon SageMaker Savings Plans.
This blog will focus on Compute Savings Plans (CSPs). We’ll discuss what services are eligible for Compute Savings Plans, the financial benefits and considerations for using CSPs.
Understanding AWS Compute Savings Plans
AWS Compute Savings Plans is a flexible pricing model that offers discounts of up to 66% when you commit to a specific amount of compute spend over a set period, typically one to three years. Thus, it’s a best practice to purchase these plans when you project a consistent amount of usage over the long term.
These plans differ from Reserved Instances in that they’re not tied to one specific Instance but instead have a broad application. This flexibility allows users to maximize savings across multiple covered configurations and easily adjust workloads without changing commitments.
Which AWS Services Are Eligible for a Compute Savings Plan?
Compute Savings Plans automatically apply to the following AWS compute usage:
- Amazon EC2: These plans apply to EC2 Instances of any family, instance size, AWS region, OS (operating system), or tenancy. For example, you can switch usage from U.S. West 1 (Northern California) to U.S. West 2 (Oregon) or from Linux to Windows and even change instance types, like M7a to M6i.
- AWS Fargate: With Fargate, the discounts don’t apply to all dimensions — only memory (GB per hour) and compute (vCPU per hour).
- AWS Lambda: They also apply to Duration and Provisioned Concurrency for Lambda function invocations at Savings Plans rates of up to 17%.
They also cover Dedicated Hosts and On-Demand Capacity Reservations.
Compute Savings Plans offer one- and three-year commitment options, with three-year terms providing higher discounts for EC2 and Fargate, while Lambda does not gain additional savings from three-year commitments. This allows businesses to choose the term that best fits their usage patterns and maximize cost efficiency.
The Financial Benefits of an AWS Compute Savings Plan
Apart from optimizing cloud spending, organizations can look forward to the following financial benefits with AWS Compute Savings Plans:
Spend-based flexibility
One of the most powerful advantages of AWS Savings Plans is their horizontal flexibility. Unlike RIs, which are restricted to specific instances, Savings Plans allow you to commit to a certain spend measured in dollars/hour. In other words, you’re not locked into specific resource configurations. Instead, your organization can change instance types, regions, etc., giving you the flexibility to adapt as needed while still realizing savings.
Matching usage
Savings Plans are designed to automatically apply to the highest percentage savings opportunities first. This ensures that you always get the maximum benefit possible from your commitment. If there are multiple instances where the savings percentage is the same, the Savings Plan is then applied to the usage with the lowest rate for that service. This method of applying discounts ensures that your cost optimization efforts are as efficient as possible.
Automatic application
With CSPs, you can change your cloud infrastructure as your needs evolve and still maintain your savings without needing to modify or exchange your commitment. Unlike traditional reservation models, where any infrastructure change might require modifying or exchanging a commitment, Savings Plans automatically adjust to your new infrastructure usage.
AWS, for example, handles this through its automatic allocation logic, which means you don’t have to manually reassign or adjust your commitments when you change the instance types or switch between services. By applying discounts automatically to eligible services, AWS Savings Plans reduce complexity and simplify cost management, giving your organization more effortless, streamlined control over cloud costs. In turn, this centralized approach makes it easier for you to predict cloud spending, months and even years down the line.
Flexibility in payment options
Compute Savings Plans have three payment options, each with different levels of discounts, commitment terms, and flexibility.
- No upfront: You make no upfront payment, and your commitment is charged monthly. This payment option offers the lowest savings. It’s suitable if you need to avoid tying up large amounts of capital to maintain liquidity. AWS also allows you to queue payments to cover a new workload you anticipate in the future.
- Partial upfront: You pay half of your commitment upfront and clear the balance in subsequent monthly payments. This payment option offers higher discounts and is ideal if your AWS costs require a significant investment and it makes more sense to spread out the payments.
- All upfront: You pay for the entire commitment in one full payment. This payment option offers the highest savings and is ideal if you run large-scale applications, are looking to minimize ongoing compute costs, and are comfortable with tying up large amounts of capital.
The Considerations of an AWS Compute Savings Plan
There’s a lot you should consider when choosing AWS Compute Savings Plans to make a well-informed decision that aligns with your business needs and financial goals, including:
Commitment lock-in
When you commit to a CSP, you agree to a fixed term, typically 1 or 3 years of a certain usage level or spend. While this ensures significant savings, it also locks your organization into a predefined level of cloud usage for the duration of the term.
This commitment lock-in risk can be problematic if your business needs change or your cloud usage decreases. In such cases, you may find yourself paying for unused resources, as the committed spend or usage cannot be reduced or refunded. While you still retain some flexibility in changing infrastructure types, the commitment to the financial spend remains rigid, potentially leading to wasted budgets if projections are not accurate.
And forecasting is challenging, especially if you have a dynamic workload that requires multiple Instances or the flexibility to change workloads or move between clouds.
You may find yourself locked into a plan that doesn’t fully support your needs. When demand increases, you have to purchase additional commitments, and when there’s a slump, you must still pay for the unused hours.
Complexity in forecasting
To maximize cost savings with CSPs, you must commit to a consistent amount of usage over the long term. However, with a dynamic workoad, forecasting potential shifts in usage can be complex when you apply the discounts to different services or to Instances of varying regions.
Risk of over- or under-provisioning
Underprovisioning would lead to budget overruns and having to purchase more plans, often at significantly higher rates. And overprovisioning can lead to wasted funds because you pay for more commitments than you need. These underutilized hours go to waste, reducing the benefit of the discounts.
Managing excess commitments can complicate your architecture, making it harder to optimize usage. It’s no wonder businesses are usually nervous about committing and tend to get stuck in the under-provisioning loop, negating the cost savings of the plans.
The trick is to monitor and analyze usage patterns to accurately predict workload demands and then balance provisioning while ensuring adequate resources.
Administrative overhead
Leveraging Compute Savings Plans involves plenty of optimization work, including monitoring, reporting, analyzing usage patterns, and cost allocation. You must also manage the various commitment terms and renewals of the CSP, and your team may need training to learn how to effectively utilize and manage the plans.
All this requires time and resources that contribute to administrative overhead. But ProsperOps can help you alleviate this burden with autonomous optimization to maximize your discounts and usage.
To replicate ProsperOps, you would need 24/7 management by an expert in both cloud discount instruments and mathematical optimization (e.g., complex linear programming), as well as detailed knowledge of the cloud systems that govern discount instruments. This would require a huge investment. But, we do it effortlessly and fully autonomously, enabling you to save time and effort for more business-driven processes.
Opportunity costs
Since you commit to a specific hourly rate, your capital may get tied up in a plan, even if your usage decreases below the commitment level. Therefore, you might not have the funds available to allocate to other projects or resources that are of higher priority or could potentially provide better returns.
How Do You Choose Between Compute Savings Plans and EC2 Instance Savings Plans?
The choice between Compute Savings Plans and EC2 Instance Savings Plans largely depends on your workload characteristics, risk tolerance, and financial goals. But, the key factor is the level of flexibility you need in your infrastructure.
- Compute Savings Plans: These are ideal if you need greater flexibility in how you use compute resources. With Compute Savings Plans, you can switch between EC2 instance families, sizes, regions, operating systems, or even shift workloads to Fargate or Lambda, without losing your savings. This option is perfect for businesses that expect their infrastructure needs to change over time or want the freedom to adapt quickly.
- EC2 Instance Savings Plans: This plan is more specific, locking savings to particular instance families and regions. While it provides savings for a specific usage pattern, it’s less flexible than the Compute Savings Plan. However, EC2 Instance Savings Plans typically offer slightly higher discounts for organizations with predictable, steady workloads that don’t require much adjustment.
In summary, choose Compute Savings Plans if you need flexibility across different compute services, and opt for EC2 Instance Savings Plans if your usage is stable and predictable, allowing you to benefit from higher savings for specific instances.
What Reports Does Amazon Provide To Help Manage Savings Plans?
AWS provides coverage reports and utilization reports to help you manage Savings Plans. With the coverage report, you can select a period and see how much of your eligible spend is being covered by AWS Savings Plans within that period.
The utilization report shows what percentage of your Savings Plans commitment you’re utilizing at an hourly, daily, or monthly granularity, based on your On-Demand usage over the selected time period.
However, these reports have their limitations and may not provide a complete picture of overall cost or enable you to optimize costs. For example, the utilization report focuses on the percentage of your AWS usage your plan covers, which doesn’t provide sufficient granularity or visibility.
Thus, you need to keep a check on one crucial aspect of commitment management: your Effective Savings Rate. ESR combines utilization, coverage, and discount percentages into a single metric, giving you the most accurate measure of how well your commitment plans are performing. By using ESR, you can assess the true ROI of your rate optimization efforts and benchmark savings across departments.
How do you calculate ESR?
This formula helps you evaluate how much you’re saving with discounted rates relative to what you would be paying on demand. In addition to being a single ROI metric, it helps you benchmark and track cost optimization efforts and offers insight into how effectively commitments are managed.
Optimize Your AWS Spending With ProsperOps
There’s a lot to consider when choosing Savings Plans — and that’s just the first step. Management complexities, changing architectures, volatile AWS usage trends, rapidly evolving technologies, lock-in risks, and other factors make optimizing Savings Plans a complex job.
That’s why you need a FinOps platform that can automate complex and time-consuming rate optimization tasks and deliver better outcomes.
ProsperOps delivers cloud savings-as-a-service, automatically blending discount instruments to maximize your savings while lowering commitment risk. Using ProsperOps’ autonomous discount management, we optimize the hyperscaler’s native discount instruments to reduce your cloud spend and place you in the 98th percentile of FinOps teams.
Using machine learning algorithms and advanced data analytics, ProsperOps can continuously analyze your company’s AWS usage patterns to identify inefficiencies and autonomously manage the commitment-based discounts based on the purchase option (all upfront, partial upfront, or no upfront) or term (one or three years) you select.
This hands-free approach to AWS cost optimization can save your team valuable time while ensuring automation continually optimizes your AWS discounts for maximum cost savings.
Schedule a live demo to learn how you can return the most savings to your bottom line with our automated FinOps platform.
