Tag Archives: Reserved Instance

Azure savings plan for compute

Microsoft announce the Azure savings plan for compute. This enables customers to take advantage of better pricing across a range of Azure compute services by committing to spend a fixed hourly amount for one or three years. Once the plan is purchased the lower prices are automatically applied to eligible services across all Azure regions. However, Reserved Instances still offer the best prices for customers who are committed to a specific VM family in a specific region.

As a comparison, a D8s v4 VM running in the East US data centre has a PAYG price of $0.7520 per hour, a 1-year RI price of $0.5938 per hour (21% savings), and a 1-year savings plan price of $0.6314 (16% savings).

This is a useful page to get started on the Azure savings plan for compute: https://bit.ly/3gipyJp, and this is where you’ll find the detail on how to buy a plan and how discounts are applied etc: https://bit.ly/3VDy1XO.

General Availability of Capacity Reservations

Microsoft announce the General Availability of Capacity Reservations for virtual machines.

A Capacity Reservation is linked to a specific virtual machine type running in a specific Azure data centre and ensures that there is always capacity available when you want to run a virtual machine of that type in that data centre.

They are priced the same as the Pay-As-You-Go prices for the VM, and if you have a matching Reserved Instance, that will automatically be applied to cover the charges of the Capacity Reservation.

Find the announcement here: https://bit.ly/3kiL53J, and a video explaining this new option here: https://bit.ly/3LFqVNr.

5-year Reservation available for the Azure HBv2 virtual machine

Microsoft announce that a 5-year Reservation is now available for the Azure HBv2 virtual machine. The existing 1 and 3-year Reservations give an estimated 25% and 50% saving on the compute costs of the VM, and the new 5-year Reservation increases savings to approximately 67%.

The announcement (https://bit.ly/32EWfI9) states that terms and conditions are mostly the same as for all other types of Reserved Instances, and page 54 of the July 2020 Product Terms confirms that the early termination fee for a 5-year Reservation is 35%, rather than the usual 12%.

Azure Reservations

Azure Reservations help customers to save money when they can make a duration commitment to an Azure service. Typically the duration is one or three years with, of course, the best savings to be made on a three-year commitment.

Today there are sixteen categories of Reservations ranging from Reserved Instances for virtual machines to Reserved Capacity for storage or database services, as well as Software Plans for Red Hat or SUSE Linux. Buying and managing Reservations, and understanding how discounts are applied can be confusing, but there’s some excellent Microsoft documentation available.

Find it here to dip into as you need to, or save it for the weekend and read it end-to-end: http://bit.ly/36W4KgJ.

Windows Virtual Desktop Public Preview

Microsoft announce that Windows Virtual Desktop is now in public preview.

This new Azure service will allow customers to run Windows 7, Windows 10 or Windows Server desktops and provide free Extended Security Updates for customers choosing Windows 7.

Customers will already be licensed for the client desktops if they have Microsoft 365 F1/E3/E5, Windows 10 E3/E5 or Windows VDA licences, and for Windows Server desktops if they’ve got RDS CALs. Reserved Instances may also be used to optimise costs for the infrastructure.

This site (http://bit.ly/2HPbqo2) has all the information as well as a video that gives an excellent overview of the service.

Understanding Reserved Instance Pre-Payment

If you’re running full-time virtual machines in Azure then pre-paying for the compute power via an Azure Reserved Instance is a good way of saving money. The pre-paid amount is automatically applied to your running virtual machines and if you want to find out more about how this (apparently magic) process works, then this article is worth a read: http://bit.ly/2HKvLsA.