The Challenges of RI and Savings Plans (part 1)

The challenges of using AWS Reserved Instances (RI) and Savings Plans (read more about Savings Plans). Both pricing models for EC2, RDS and other AWS services.

The core principle is simple – commit to a minimum value of use, and enjoy up to 75% off AWS cloud services. Enterprises that care about the costs of their cloud infrastructure make use of the reserved instances and the Saving Plan to reduce costs over the long term and stabilize the prediction of infrastructure workloads.

However, the management of these charging models is not as simple as it may seem. Because they are only a billing mechanism, the terms and conditions of each pricing model can easily vary. Misinterpretation, caused in large part by the complexity of managing the reserved instances and the Saving Plan. It can end up generating unnecessary expenses or loss of savings opportunities.

To understand the various parameters that must be considered when purchasing reserved instances, one must initially understand the various types of reserved instances offered by AWS.  The types of reserved instances vary according to different characteristics. Which affects their price and their respective savings potential.

1/ Characteristic of reserved instances (RI)

• Types of Commitment: The purchase of reserved instances of AWS may involve a commitment of 1 or 3 years.

• Scope of Territory: Reserved instances may be regional or limited to a specific availability zone within a region (zonal).

• Type of instance : The reserved instances correspond to a specific family group.

• Offer Classes (Standard / Convertible): While some assignments of the reserved instances can be modified, the family of a Standard instance cannot be changed. However, when using convertible reserved instances, the family group of the instances can be exchanged for another type.

• Payment Options: Reserved Instances can be paid “fully upfront”, partially upfront, or without any upfront payment.

• Platform: The reserved instances, as well as the workloads covered by them, vary according to the operating system. Example: Windows or Linux/Unix.

• Flexibility: Regional reserved instances can be flexible and are automatically applicable to different sizes of instances within the same family.

• Services: There are different types of RIs for different types of AWS services. For example: EC2, RDS, ElastiCache among others.To illustrate the effects that these characteristics produce. Observe below the table that shows how the savings percentages of the reserved instances vary according to the commitment term, the offer classes and the payment options.

2/ Take-away for users of reserved instances (RI)

Taking all these parameters into consideration, several points become critical for users of reserved instances.

• While prices in some instances tend to decline (as has been the case with AWS by reducing costs frequently). Prices in Reserved Instances remain fixed and related to the On-Demand prices of the purchase period.

• In addition, as new types of instances are frequently released by AWS. The reserve can be applied to types of instances with obsolete technology.

• Add to this the fact that the client is being charged by the reserved instance regardless of its use or not. With this, the saving potential of a reserved instance is only reached in its maximum when the instances are active and in use.

• Finally, while modern workloads are flexible and scalable (to meet varying use and demand), reserved instances are fixed and do not vary.

As a result, efficient management of private instances is no easy task.

3/ What’s coming next?

Reserved instances are not as directly applicable as they appear, and to ensure excellent savings as well as wide coverage of reserved instances, decisions must be made per second. You then have to decide in a short time the type, group, payment options, and other variables.

Even having extensive knowledge about the characteristics of the instances. Many consumers are having difficulties in planning and executing a strategy of buying and selling reserved instances. It takes time, and a great effort from a technical team, to face the challenge of using reserved instances on their own.

In order to solve the Challenges of Using RI and Savings Plans. The Spotinst launched the Spotinst ECO – Autonomous Manager of Reserved Instances. Learn more about Spotinst ECO in the article Spotinst ECO. An Autonomous Management of Reserved Instances and Saving Plan at AWS.

Part 2: coming soon