A couple days ago I wrote a post to compare what Google Cloud offers, but AWS currently does not support. One of them is that Google charges VM usage per-minute, but AWS charges per-hour. But now the game is changing, as AWS just announced today that ‘Effective October 2nd, usage of Linux instances that are launched in On-Demand, Reserved, and Spot form will be billed in one-second increments. Similarly, provisioned storage for EBS volumes will be billed in one-second increments.’
I am not sure if per-second is over kill. For most cases, I think per-minute should be enough, as virtual machine normally even takes more than 1 minute to boot up or shutdown. In some scenarios, it may make a differences. Container take a few seconds to boot up, may be per-second charge is designed for container related usage? Like tons of thoughts containers launched for some sort of processing jobs, and they can complete the jobs in a very fast fashion which can be counted in seconds.
Reference: https://aws.amazon.com/blogs/aws/new-per-second-billing-for-ec2-instances-and-ebs-volumes/
Jackie Chen's IT Workshop 