Amazon Web Services has changed the way reserved instances can be managed to allow more flexibility, and cloud users hope it's only the beginning of new options for the instance type.
The ability to move a reserved instance among availability zones was especially appealing to Glenn Grant, chief executive officer of G2 Technology Group, a Boston-based AWS consulting partner.
"We can come into a situation where … a customer may have purchased quite a few reserved instances in a single availability zone," he said. "Now, as consultants, we can help customers get into a high-availability architecture without throwing those [reserved instances] out."
AWS users said they'd like to see more reserved instance flexibility.
"It would be good for [AWS] to eventually allow multiregion as well," said Kent Langley, CEO of Ekho Inc., a Web-based data analytics company based in San Rafael, Calif. "When I deal with AWS, I like to think of all the regions as my meta data center -- I don't like it when there are artificial constraints that keep me from being able to be fluid in my use of the multiple global regions."
One commenter on an AWS blog post about the new reserved instance features also suggested that the ability to resize the instances within a certain total pool of resources be added as well.
To a certain extent, reserved instances won't ever be as flexible as on-demand instances because AWS relies on them for capacity planning and business forecasting, said Nicolas Fonrose, founder of Teevity, a cloud computing monitoring software startup based in France, in a separate interview.
There's some disconnect between the reserved instance business model and what users might ultimately want -- the ability to specify just the AWS region and amount of compute power they plan to burn in the coming month or years, as well as the level of intensity. Still, "I'm sure there is room between what AWS offers today … and what users would dream of," Fonrose said.
Not every AWS customer agrees.
"In this case I think there's a middle ground we'll have to come to with AWS," said Joey Imbasciano, cloud platform engineer for Stackdriver Inc., a cloud monitoring service company based in Boston. "Just as we need to capacity-plan, AWS also uses [reserved instances] for planning purposes. I think a feature like this would make it very challenging for them to plan."
Currently, users who wish to change the capacity on reserved instances can also use the spot instance marketplace to resell unused capacity to other AWS customers. Pricing for reserved instances remains the same; pricing for a small instance on a one-year term with light utilization starts at 3 cents per hour, with a $61 up-front fee.
An Amazon spokesperson said the company welcomes customer feedback on its services.