Introduction
Reserved Instances (RI) for Azure are similar to AWS Reserved instances in various ways and yet have their own set of differences, offering a great amount of savings and flexibility depending on the requirements at a given point in time. Comparing the pricing between Azure and AWS cloud has always been a challenge with frequent fluctuations in prices but Azure Reserved Instances are known to have been able to offer higher cost benefits than the on-demand prices of AWS Reserved Instances. However, Azure RI come with a flipside and in this post, we will have a look at what exactly Azure Reserved Instances are and how similar or different they are in comparison to AWS Reserved Instances.
Azure Reserved Instances
Microsoft introduced Azure Reserved Instances (Azure RI) to enable their most active Azure customers to achieve long-term savings from VM usage. By promising a one-year or three-year benefit, Azure customers will be able to pre-book VMs at significant discounts, saving up to 82% over standard pay-as-you-go models. Customers who have somewhat predictable workloads and resource needs, benefit most from Azure RI, also enabling them to set budgets and make long-term plans.
Despite the need to commit to an upfront payment, Azure RI is not confined by a rigid, long-term framework. With Azure RI’s flexibility, users can continuously modify or terminate services. For instance, an enterprise can replace one RI with another RI of a different VM type or with one that runs in a different region in case of a change in the compute demand in a given time period. Here, customers can purchase a new RI through the prorated refund received from early termination, turning Azure RI into a very cost-effective alternative for consumer-based cloud computing.
- Azure accounts and subscriptions: While Azure accounts are similar to AWS master payer accounts, Azure subscriptions are similar to AWS sub-accounts. Azure accounts are used for general billing purposes, but each subscription generates its own set of billing data helping you allocate costs. In addition, each subscription can be used for access control and environmental isolation.
- Scope: Reserved Instances can be “single” and applied to VMs (or other services) for a particular subscription, or “shared” to apply to reservations for other subscriptions they own or have the same billing context as the reservation.
- Flexibility: When you purchase an Azure Reserved Instance, you can select “Instance Size Flexibility”. This ensures that reservations are applied to VMs of any size in the flexibility group of the same instance size. For example, if you purchased a reservation for Standard_DS4_v2 with a footprint of 8, but you were running two VMs with a footprint of 2 and a Standard_DS2_v2 size along with a VM with a footprint of 4, the first purchase is fully booked covering the cost of the actual consumption.
- Preferred capacity: For Reserved Instances with a single scope, you can select the capacity of your preference. This, as the name implies, prioritizes the data center capacity of your deployment. This option is similar to AWS Zone Reserved Instances.
- Target services: Azure provides a detailed list of locations that reservations can be applied to. In addition to Azure VMs, eligible services include Blob storage capacity, Azure Database for MariaDB, MySQL, PostgreSQL compute components, and other Azure services.
- Reservation cancellation: When you no longer need a reserved instance you purchased, you can redeem it for another instance family or region or even cancel reservations worth up to $ 50,000 without penalty.
- The benefit of Azure Hybrid: For customers with on-premises Windows Server and SQL Server licenses, you can realize cost savings of up to 80% when running in Azure with RIs.
A Detailed Comparison Between AWS & Azure
Despite having numerous similarities, AWS and Azure Reserved Instances differ in a few significant ways. So, let’s have a look at some of their distinctive qualities:
AWS
With a wide range of purchase options, AWS Reserved Instances are a more established and sophisticated product. These offer clients the option of:
- Long-term, fixed commitments in exchange for greater discounts
- More flexibility but less potential savings
In terms of flexibility, providers classify reservations as either Standard or Convertible Reserved Instances. It also provides additional types of reservations, scheduled reservation instances, and engineered reservations.
Standard Reserved Instances
Generally, the higher percentages of savings are available on Standard Reserved Instances. However, once you’ve made your reservation’s specifics known at the time of purchase, they’re essentially fixed in stone.
The standard version cannot be switched for a different instance family, operating system (OS), tenancy, or payment method, although regionally scoped reservations can be switched from one Availability Zone to another, and vice versa. Additionally, you can assign a reservation to a different Availability Zone within the same region, as well as change between the EC2-VPC and EC2-Classic network types.
Ultimately, if your operating system is a free version of Linux, you can change the size of your instance and combine and divide reservations with the same start/end times.
Convertible Reserved Instances
Standard reservations can be replaced with convertible reserved instances for greater flexibility. However, there are costs associated with this, and the highest savings are just around 45%.
As long as the exchange is equivalent to or superior to the current one, you can change your instance family, operating system, or tenancy. If you have ongoing IT projects or too much reserved capacity, this offers a safety net.
Scheduled Reserved Instances
Scheduled Reserved Instances only purchase capacity for regular time slots, not continuously throughout the reservation period. This makes it a cost-effective way to host your application with typical workload patterns such as: B. Peak activity during work hours and quiet times in the evenings and weekends.
Scheduled Reserved Instances are only available for one year and must be reserved for one year. At least 1,200 hours. Currently only supported on a limited number of instance families — C3, C4, C5, M4, R3.
AWS Reserved Instances vs AWS Saving Plans
In 2019, AWS introduced Savings Plans (EC2 Savings Plans and Compute Savings Plans). With this new committed capacity pricing plan, the customer promises to pay them as per their wish on an hourly basis (e.g., $35 per hour for 1 or 3 years). In this example, all expenses up to $35 are calculated according to the savings plan (66-72% savings) and any cost above the committed amount will be billed at an on-demand rate.
However, each option has its strengths and weaknesses and therefore it is important to analyze and consider the best pricing model before choosing the final option.
- While savings plans cannot, standard Reserved Instances can be bought and sold on AWS Marketplace which provides better flexibility, ensuring you keep up to your defined committed level of spending.
- Savings plans apply only to EC2, Fargate, and Lambda, while Reserved Instances are broadly applicable to EC2, RDS, Redshift, ElastiCache, and other AWS services.
- Convertible Reserved Instances allow you to add new reservations to cover a larger number of EC2 instances during the term of your contract without having to extend the term of the contract. This is especially useful if your needs change and it doesn’t make sense to set a new term of 1-3 years.
- For savings plans, each modification to the original contract will be made in the new contract starting on day 0. EC2 Instance Savings Plan is applicable for use across all given instance families, irrespective of the operating system or tenant. Standard Reserved Instances can also be applied for use across a particular instance type, but the instance needs to be Linux and default tenancy.
- Reserved convertible instances are limited to specific instance types, OS tenants, and regions, but compute saving plans apply to all usage types across multiple regions.
Side-by-Side Comparison of AWS and Azure Reserved Instances
| AWS Standard | AWS Convertible | Azure | |
| Term | 1 or 3 Years | 1 or 3 Years | 1 or 3 Years |
| Payment options | Upfront, Partial, No Upfront | Upfront, Partial, No Upfront | Upfront only |
| Reserved Instance sharing | Yes | Yes | Yes |
| Exchange offers | Availability Zone, instance size and networking type only | Full flexibility | Full flexibility |
| Cancellation | No – you can try to sell on the marketplace | No – you can try to sell on the marketplace | Yes – 12% fee |
| 1-Year Discount | 22% to 51% | 11% to 44% | 8% to 47% |
| 3-Year Discount | 45% to 72% | 31% to 66% | 13% to 72% |
| Priority access to spare capacity | Only if assigned to an Availability Zone | Only if assigned to an Availability Zone | Yes |
| Allocating RIs | Yes | Yes | Yes |
Conclusion
Thus, the new Azure RI has some interesting advantages when compared to AWS RI. All Azure RIs are exchangeable and cancelable, subject to a fee of 12% of the RI’s residual value. AWS offers similar flexibility when exchanging RIs for its convertible offerings, but the ability to find buyers for unwanted AWS RI through the marketplace is uncertain and highly unlikely.
Azure RI lags behind AWS RI with respect to payment terms as it asks for a full upfront payment. Additionally, the “random” method of allocating Azure RIs within or across subscriptions makes it difficult for large organizations to accurately allocate both upfront costs and savings. Ultimately, whether you use AWS or Azure, purchasing Reserved Instances is an important part of your cost-savings strategy.