Microsoft Azure Reserved Virtual Machine Instances (RI's) - Key Features and Benefits

Azure Reserved Virtual Machine Instances have emerged as a popular and cost effective alternative to pay-as-you-go cloud computing services. Reserved Instances, simply known as RI's are virtual machines on public cloud network, allocated to users for a particular period of time. It is different from on-demand purchases as time period for RI's is long term, usually spanning between 1-3 years. The customer pays for the entire duration and the hourly rate is significantly lower when compared to the on-demand, pay-per-use systems. Another way to visualize RI’s would be as ‘billing discount applied to on-demand instances’. In this sense, they aren't physical instances.

If used wisely, Azure Reserved VM Instances have the potential to save huge costs for organizations.

Microsoft Azure Reserved VM Instances

Microsoft Azure, the global Cloud leader offers RI's on a variety of Virtual machine (VM's) types and customers holding the Microsoft Enterprise Agreement (EA) can apply their Azure Monetary Commitment while procuring the RI's. Offered for new customers as well, Reserved Instances can be purchased by selecting the time frame, region and then paying upfront on the Azure website. Microsoft offers up to 72% of discount on RI's when compared to pay-per-use Virtual machines pricing model.

Despite the front-up commitment, Azure RI doesn't lock the users by offering them flexibilities and adjustments on purchase plans. Users can even cancel the service by paying a small penalty or simply exchange without any fees.

Let us analyze the attributes of Azure Reserved Instances (RI's) and further discuss its benefits, usage and applications.

1) Microsoft Azure Reserved Instances (RI's) are available on all Azure Virtual Machine types, except the A-series, A_v2 series and G-series.

2) We mentioned that organizations can save up to 72% on Azure RI's as compared to pay-per-use VM’s. With the added value of Azure Hybrid benefit, organizations can even save up to 80-82%. This makes RI's a very cost efficient and sound investment.

Image Credit: azure.microsoft.com

3) RI's drive automation by managing the systems independently and allows the user to spend more time on strategy and in driving business growth.

4) Azure RI's are offered on both Windows and Linux Virtual Machines.

5) Azure RI offerings are very flexible. For example, total cost of up-front and monthly reservations are same and customers don't have to pay extra fees when they chose the monthly payment option. In addition, Azure allows users to cancel any reservation costing up to 50,000 USD annually. Users can get a refund of their unused instance by paying a termination fee of around 12 percent.

6) For the unused portion of the committed funds, users can also apply to a new Azure Reserved Instance plan of either one or three years. For this exchange, there is no fees charged.

7) Customers already on Windows servers can avail more savings.

8) One of other benefit of RI is the ability to reserve the instance in a chosen availability zone. This can be particularly useful if the user’s infrastructure in a specific Availability zone and if the infrastructure uses auto scaling and experiences frequent spikes in usage.

9) With Azure RI, users are failsafe when it comes to outages.

10) Azure RI assignments are very easy to change post-purchase.

11) Azure RI offers the full benefit of prioritized capacity. This ensures greater availability and better performance of workloads and applications in both global and public Azure regions.

Purchasing Reserved Instances.

Azure RI purchasing process is simple and streamlined. A subscription owner in an Enterprise (MS-AZR-0017P or MS-AZR-0148P) or Pay-as-you-go subscription (MS-AZR-0003P or MS-AZR-0023P) can make a RI purchase. Enterprise Agreement (EA) customers can utilize their Azure monetary commitment to purchase the instances and this monetary commitment is used first to purchase the instances. If the credits aren't enough then payment is adjusted in the next overage bill.

EA customers also have the option to limit purchases. They can go to the EA portal and disable the Add Reserved Instances option. This option comes handy where different cost centers are going to use RI's. A centralized team can make the purchase up-front and they can then just add these cost centers online. In this case, the centralized team needn't be the owners of subscription access.

Customers can also buy it via Azure.com by charging their credit card. The same medium will be charged for the next payment cycle and in case the user is outside USA, the new exchange rate against the US Dollar will be levied.

Scope Reservations

1) Single resource group scope: In this scope option, the discount is offered only to selected resources in the applicable resource group.

2) Single subscription scope: In this scope option, the RI discount is applied to the selected resources in the applicable subscription.

3) Shared scope: In this scope option, the RI discount is applied to selected resources in the applicable subscriptions which are in the billing context.

It is implemented in the following order:

• • RI, scoped to a resource group
• • Single scope RI
• • Shared scope RI's

Azure portal offers the users an easy way to monitor their RI usage. The grid on the portal displays the last recorded utilization percentage for the RI. If the user is an enterprise agreement customer, he/she can also get the feed through API's.

Recommended use

Microsoft recommends that RI's are used for workloads and applications that have predictable usage patterns. In addition, Reserved Instances (RI's) are also suited for organizations who have a long term investment in cloud.

Microsoft also helps the user to select the correct Virtual Machine(VM) size for their Azure Reserved VM instances. It analyzes the monthly usage activity and recommends saving potential plans.

In addition, Azure RI offers the flexibility of exchange across any region and for any series, depending on changes in user's applications and workloads.