Introduction to AWS Pricing to Help You Optimize Cloud Costs
Amazon Web Services offers a dizzying range of products and services for computing, storage, database, analytics, application, and deployment on its cloud platform. Understanding the costs of such a wide portfolio can be overwhelming for customers, especially because the pricing for all services is unique, and they change and evolve frequently.
In this article we will try to bring clarity to the different product and service pricing models by AWS. Our focus will mainly be on AWS pricing for compute and database services in the cloud.
For computing and database services AWS offers five types of pricing structures:
- On-demand pricing
- Spot Instances
- Reserved Instances
- Savings Plans
- Volume discounts
Let’s review each AWS pricing model to understand how they work and help you decide which one suits your business needs in the cloud.
This pay-as-you-go AWS pricing model allows organizations to pay for the services that they use on demand, as needed. Compute resources, including AWS EC2 pricing, are assessed by the hour.
On-demand pricing may make sense in the following scenarios:
- For non-production workloads, such as development, testing, QA, and training
- For companies that have recently migrated to AWS cloud. New customers need time to understand their resource utilization and establish a clear pattern of their usage in the cloud. Based on their usage patterns they can evaluate the more complex pricing models and weigh the benefits of reserving instances in advance and purchasing savings plans for potential cost reductions based through volume purchases.
On-demand pricing allows organizations to quickly adapt to changing business needs without overcommitting budgets and other resources in the cloud. As demand increases however, and a fixed number of instances are constantly running, on-demand rates can quickly add up, forcing companies to do longer term planning and forecasting for higher-volume purchases in the cloud.
Spot Instances Pricing
In this EC2 pricing model, customers can acquire unused Amazon EC2 instances for discounts of up to 90% off the on-demand price.
Spot instances may make sense for the following scenarios:
- Applications that have flexible start and end times
- Applications that are only feasible at very low computing prices
- Test and development workloads with stateless, fault-tolerant, or flexible applications, such as big data, containerized workloads, continuous software integration and deployment, web servers, high-performance computing
- Urgent computing needs for large amounts of additional cloud capacity
The key drawback of spot instances is that they can be interrupted at any time with only a 2-minute notification when EC2 needs that capacity back. Customers need to be able to handle that interruption to take advantage of spot instances.
Customers can check the spot price for each region and instance type on the AWS website, which is updated every 5 minutes.
Reserved Instance Pricing
For computing services, like Amazon EC2 pricing, and database services, like Amazon RDS pricing, you can reserve a set number of instances in advance, for a significant discount (up to 60%) compared to on-demand rates. Instances can be reserved at either one-year or three-years in advance and require upfront payments. Larger upfront payments yield greater discounts.
With Reserved Instances customers commit to usage attributes. The discounts are automatically applied when customers use an on-demand instance that matches the attributes of an active Reserved Instance.
Reserved Instances carry the following attributes:
- Instance type: Instance types consist of different combinations of CPU, memory, storage, and networking capacity. For example, m4.2xlarge.
- Platform description: RIs can be purchased for instances running on Linux/UNIX, SUSE Linux, Red Hat Enterprise Linux, Microsoft Windows Server, and Microsoft SQL Server platforms.
- Tenancy: Each EC2 instance on AWS has the following values:
- Default: Your instance runs on shared hardware.
- Dedicated: Your instance runs on single-tenant hardware.
The tenancy of a default instance cannot be changed after it is launched.
- Availability Zone (optional): RIs purchased will provide a reservation and a discount when the instance is used in a selected Availability Zone. When an Availability Zone isn’t specified, the RI discount will apply to an active instance of any size (within the same family) in the region. For example, if you own an m4.2xlarge Linux/UNIX RI with default tenancy in US East (N. Virginia), the discounted rate for this RI can automatically apply to two m4.xlarge instances in us-east-1a or four m4.large instances in us-east-1b.
Standard vs. Convertible RIs
Companies can choose the RI type that best fits their applications needs in the cloud:
- Standard RIs: These Reserved Instances allow customers to modify Availability Zone, scope, network platform, and instance size for Linux RIs.
- Convertible RIs: Customers can exchange the Convertible RI for another one with new attributes, including instance family, instance type, platform, scope, and tenancy.
RI Payment Options
There are three payment options for Reserved Instances:
- No upfront: No upfront payment is required. RIs are billed monthly. This option allows you to preserve capital to invest in other projects
- Partial upfront: A portion of the cost is paid upfront, the rest is billed at a discounted hourly rate, regardless of whether the RI is being used.
- All upfront: Full payment is made at the start of the term, with no additional hourly charges incurred for the remainder of the term, regardless of hours used. This option provides the largest savings.
The below table illustrates the cost savings available to customers based on the commitment term and the type of RI used:
Parquantix pays the upfront fees for qualified customers, helping them preserve cash and invest in other areas of their business. Contact us for a 30-minute consultation.
Reserved Instances make sense in the following use cases:
- Production environments. Most production applications require servers to be available 24×7.
- Steady state use cases with predictable workloads.
- Mission critical applications that run on Multi-Available Zone database deployments to ensure higher availability and data stability.
In addition to cost savings, RIs provide flexibility to change instance families, OS types, and tenancies with ease. Reserved Instances help you minimize risks and more predictably manage budgets.
AWS Marketplace for Reserved Instances
The Reserved Instance Marketplace allows AWS customers to list their Standard RIs for sale. This is an effective incentive for customers to offload their unused RIs before they expire, or to offload older instances when it is time to upgrade to the latest versions. It allows sellers to offset part of the cost of their instances, and allows buyers to find the instances they need at a fraction of the cost, with shorter term commitments.
One major limitation of the AWS Marketplace is that it is only available for businesses with a US bank account. However, Parquantix helps its customers that are located outside US to sell their unused instances in the AWS Marketplace, generating additional savings.
AWS charges a service fee of 12% on the total upfront price for each Reserved Instance sold on the Amazon EC2 Reserved Instances Marketplace.
Savings Plans were recently introduced to simplify AWS pricing for volume purchases. They offer significant discounts (up to 72% off, according to AWS ) when customers commit to spend a specific dollar per hour amount over a one, or three-year term. Any usage that exceeds the purchased volume of hours are billed at the regular on-demand rate.
Users can choose between two plan types:
- Compute Savings Plan (CSP): This can be applied to any instance in any region regardless of OS, instance family, size, Availability Zone, region, or tenancy. The AWS pricing for these is identical to convertible RIs. AWS has, in effect, modified convertible RIs to be used in any region without having to actually do the conversion. In addition to Amazon EC2 pricing, discounts apply to AWS Lambda and AWS Fargate usage.
- EC2 Instance Savings Plan (ISP): This plan provides discounts in EC2 pricing in exchange for a usage commitment of individual instance families in a specific region, regardless of Availability Zone, size, OS or tenancy.
Reserved Instances and Savings Plans can be used together. Savings Plan discounts apply to any on-demand usage that is not covered by RIs.
Savings Plans make sense for the following use cases:
- AWS Fargate and AWS Lambda usage, as these cannot be Reserved for RI discounts
- Customers with limited resources to monitor and optimize cloud usage
- Introduced as an alternative to Reserved Instances, Savings Plans lack the flexibility of RIs. Additionally, companies don’t have the option of deferring the capital investment for upfront payments, because Savings Plans require at least 50% of the cost to be paid at the time of purchase. For a detailed comparison of Reserved Instances and Savings Plans you can view our earlier article.
AWS Pricing with Volume-based Discounts
AWS provides volume based discounts to help you obtain more savings as your usage increases. For outbound data transfers from EC2, tiered pricing lets you pay less per GB usage. As your AWS usage needs increase, you benefit from the economies of scale that allow you to keep cloud costs under control.
In addition, if you purchase a large number of Amazon EC2 RIs, you will receive discounts on upfront fees and hourly fees for future purchases, as illustrated below. If your RI volume exceeds $10 million, Parquantix will negotiate further discounts directly with AWS.
Contact us for a consultation on how we can maximize your savings based on the optimum Reserved Instance strategy.
AWS Free Tier
An introductory free tier is offered to all new AWS customers when they sign up. The offer is valid for 12 months from the day they register on AWS. It provides 750 hours of free instance usage for Amazon EC2 and Amazon RDS, and 5GB of storage capacity with Amazon S3.
Should you take advantage of AWS Free Tier?
AWS Free Tier is the entry point for new customers in the cloud, and helps them gain hands-on experience with AWS products, and services, without the cost commitment. Certain products, such as Amazon Lambda, don’t have a 12-month limit.
In what circumstances does each pricing model make sense?
You can start with on-demand pricing to establish a clear usage pattern. For predictable EC2 and RDS resource requirements, RIs will provide the maximum savings. Savings Plans will lower the cost of AWS Lambda and Fargate.
What pricing model will save you more in the long run?
It depends on your specific business needs and usage patterns, including whether you will be running production environments or developing and testing applications; if you have predictable or variable workloads; based on your availability and scalability requirements, and the underlying OS.
What are the factors to consider in optimizing costs?
AWS pricing is complex. The pricing is further complicated by the large portfolio of AWS products and services, all with their unique pricing structures, with multiple tiers of discounts based on volume of usage, which change on a regular basis. The average customer needs a dedicated resource with a high level of involvement to monitor, learn and manually implement all the AWS pricing models to be able to optimize the cost of their unique cloud deployments.
Continually monitoring usage is necessary to control and optimize cloud costs. Tracking performance dynamically helps prevent overprovisioning, trigger actions to resize or shut down instances that are underutilized, apply reserved instances where actual usage occurs, and sell unused instances before they expire.
Fortunately, Parquantix has an automated AI solution that eliminates the time, effort and expense required to do all that. The solution’s algorithms are updated with every new product and service introduction, pricing change and technology upgrade. It monitors usage on an hourly basis, optimizes utilization and maximizes your savings in real time. As a result, you don’t have to waste critical time away from what matters most: Running and growing your business in the cloud.
Learn more about how our automated tool can help you navigate the complex AWS pricing models to optimize your cloud usage.