AWS to Start Billing Linux Instances By the Second

  • Update: Google One-Ups AWS’ Per-Second Billing Plan

Amazon Web Services (AWS), announced Monday that it will introduce per-second billing to certain Elastic Compute Cloud volumes and Elastic Block Store volumes. According to Jeff Barr, AWS evangelist, all Linux EC2 instances, including those that are on-demand, reserved, and spot, will start Oct. 2 and be billed in second-long increments. Previously, EC2 users had to pay per hour for on-demand instances, regardless of whether they used their rented compute power during that time period. “Some of our most sophisticated customers have developed systems to maximize the value of EC2 by strategically selecting the most profitable target instances for managing their gaming, adtech, or 3D rendering fleets. Barr explained in his blog that per-second billing eliminates the need for instance management and brings cost savings to all customers. He said that the per-second billing scheme will apply to provisioned storage for EBS volumes as well as Amazon Elastic MapReduce and AWS Batch services. This move is directly against AWS rivals Google Azure and Microsoft Azure. These companies have used per-minute billing for years to charge users for virtual machines (VMs), on their respective platforms. Although Microsoft did introduce a per-second billing system in July, it only applies to Azure containers and not VMs. Barr described the new per-second billing scheme as a way for organizations and developers to reduce unneeded instances usage. However, there are some caveats. This new billing scheme has a one-minute minimum per instance. Barr stated that per-second billing is not available for Windows instances and “Linux distributions which have an hourly charge” at the moment.