Amazon has tweaked to the pricing for its EC2 instances, introducing the chance for firms to reserve short-term capacity and subsequently sell on extra that it decides not to use.
The so-called Reserved Instance Marketplace allows firms to buy up fixed-term capacity at a discount. And they can now also auction off any capacity they decide not to use.
“As more and more customers have started buying Reserved Instances, they’ve asked for ways to sell their Reserved Instances to change instance types or where instances are located, and they’ve asked for more Reserved Instance term length options,” explained Peter De Santis, vice president at Amazon EC2.
Amazon's EC2 Reserved Instance Marketplace complements its on-demand spot market, which allows firms to pay for compute capability by the hour, the firm explained.
Previously, customers were limited to buying up allocations for either a one-year or three-year period. But the new changes will give customers greater options over the time frames they want.
According to Amazon, customers can expect to save up to 71 percent by choosing to reserve their allocations.
Once firms decided they no longer wish to use their full allocation, perhaps because a project has finished early, they look to sell spare capacity to others.
Those that find buyers will have their transactions brokered by Amazon, which should make the task of selling space capacity relatively painless.
The changes are perhaps to be expected: many businesses have been wary about using spot markets and prefer their IT costs to be predictable.
The updated EC2 reservation system may provide such users with a greater degree of flexibility, without being exposed to price spikes. The downside is that they have to pay up front.
Would you want to live in a world without memes?
Traditional theories debunked by new study
Scientists closer to developing material capable of splitting water for better storage of solar energy
Experiments needed to see if the material works in the real world
Developers first in the queue to test TensorRT and TensorFlow integration tools running on Nvidia GPUs