AWS Releases Next-Gen ‘Burstable’ Instances

Four years after introducing the T2 instance type, Amazon Web Services Tuesday brought online the T3s, providing customers with volatile workloads greater savings potential.

The T3s, available to provision in 12 AWS regions around the world, "are even more cost-effective than the T2 instances," AWS chief evangelist Jeff Barr said in a blog post.

The newest EC2 virtual machines improve on their predecessors in price performance by 30 percent, according to Seattle-based AWS.

[Related: AWS Debuts 4TB EC2 Instances For Powering Memory-Intensive SAP Workloads]

The T-series makes possible cost savings for workloads that typically don’t approach the maximum performance of their CPUs, but do require periodic bursts of computing power. AWS delivers the baseline percentage of processing power, and customers accumulate credits when idle that can be consumed later. Those credits are stored for up to seven days.

Last year, AWS bulked up the T2 instances with an option for unlimited bursting capacity to handle spikes in demand. With that plan, users pay for bursting when their average usage exceeds the set baseline threshold of computational power over 24 hours. The T3s comes with unlimited bursting by default.

On-demand prices for the T3s start at $0.0052 per hour, or $3.796 per month. If the instance exceeds the baseline, AWS charges 5 cents per vCPU-hour.

"This is a unique form of vertical scaling that could very well enable some new types of applications," Barr said in the blog.

The T2s has proven popular for hosting microservices, low-latency interactive applications, virtual desktops, build and staging environments and prototypes, Barr said in an earlier post.

Barr suggests AWS customers running M4 or M5 instances who don't actually need sustained compute power consider transferring those workloads to T3s.

"You can host the workloads at a very low cost while still having access to sustainable high performance when needed," he said.

The T-series delivers a "generous and assured baseline amount of processing power," Barr noted, with the ability to scale up to the full core (2.5 GHz Intel Xeon scalable processors) when more processing is needed, and for as long as it's needed.

The series is available in seven sizes, from “nano” to “2xlarge.”