Amazon Adds 'Scheduled' Reserved Instances To Server ListAmazon Adds 'Scheduled' Reserved Instances To Server List

6 Machine Learning, AI, Analytics Trends To Watch

Amazon Web Services has launched a variation on its Reserved Instance class of virtual servers with what the company calls "Scheduled Reserved Instances." By designating a time when the Reserved Instance will be used, customers may gain an additional discount of 5% to 10% on Amazon's cheapest cloud-computing option.

Scheduled Reserved Instances, unlike regular Reserved Instances, designate virtual servers that will be repeatedly used at specific times of the day, week, or month and are paid for only for use during those times. Regular Reserved Instances (now called Standard Reserved Instances by AWS), on the other hand, are guaranteed to be available 24 hours a day throughout the period of the contract.

By being specific about use times (and restricted to them), customers gain an additional discount with the Scheduled version.

Both types are in contrast to On-Demand instances, which are fired up without prior notice upon receipt of a customer request and paid for by the hour. When a customer gives Amazon a one- to three-year commitment for Reserved Instance use, Amazon is willing to discount the server by up to 75%, according to information on its website. Scheduled Reserved Instances, therefore, might be as much as 80% to 85% off the price of a comparable on-demand instance.

They may be scheduled for a one-year period, according to a Jan. 13 blog post written by Jeff Barr, chief evangelist for AWS.

Customers initially were required to make an upfront, down payment on for Reserved Instance use. However, Amazon has since offered more flexible ways to buy Reserved Instances with only a small down payment or no down payment. Regular Reserved Instances are guaranteed to be available at any time during the period that the customer has agreed to be billed for it.

In his blog post, Barr wrote that the Scheduled Reserve Instances would be good for a bank or mutual fund that performs risk calculations each day in the afternoon, or a phone company that runs a multi-day billing application at the start of each month. Another example would be a trucking company that seeks to optimize shipment routes each Monday, Wednesday, and Friday morning.

"Our new Scheduled Reserved Instances are a great fit for use cases of this type. They allow you to reserve capacity on a recurring basis with a daily, weekly, or monthly schedule over the course of a one-year term," Barr wrote.

Scheduled Reserved Instances now appear on the Amazon EC2 management console, with the customer selecting the virtual server resources desired and designating the time they will be used.

Bill Lynch, vice president of product at Cloudability, a cloud usage optimization service, noted in a blog that the time will be designated in Universal Coordinated Time (UTC), which is the equivalent to the UK's Greenwich Mean Time, but tricky to calculate for those not familiar with the time difference.

There are UTC converters to your location available online, including this website.

[Read why Amazon cut cloud prices to start the year.]

Lynch said that Scheduled Reserved Instances would be good for a development team's code builds and testing. And he pointed out in his blog an important detail that was left out of Barr's introduction to the instance type: There are minimum use requirements.

"Depending on the type of schedule (daily, weekly, monthly), you're committing to a minimum of 4 hours per day, 24 hours a week, or 100 hours a month of compute for an entire year," Lynch wrote.

Something else to remember: Unlike other Amazon virtual server types, Reserved Instances and Scheduled Reserved Instances run in a specified availability zone in a particular region and can't easily be switched from one place to another.

Customers may let Amazon make the assignment, but once made, the Reserved Instance will stay there. Amazon appears to require certainty of location for both types of reserved instances in order to map reservations for capacity to specific physical data centers.