AWS T3 Vs. M5 Instances: A Side-By-Side Comparison

While T3 and M5 instances aren’t the newest ones available on AWS today, they aren’t the oldest either. We compare Amazon EC2 T3 and M5 instances today – and highlight the key differences so you can choose the right EC2 instance types for your workload requirements.

What Are Amazon EC2 T3 Instances?

T3 instances are the third generation of Amazon EC2 T family of instances. Amazon EC2 T3 instances fall under the General Compute instance class. T3 instances offer balanced performance in terms of compute, memory, and networking.

Unlike T2 instances, T3 instances have Burstable Mode enabled by default. With this feature, T3 instances will perform at baseline capacity until the load increases, at which point they will burst CPU usage.

Burstable mode enables T3 instances to run at moderate CPU usage most of the time, but to leverage 100% of a CPU core for one minute, or for as long as necessary below the 100% level, improving performance under heavy load.

On the other hand, a T3 instance accumulates credits while it runs at or below baseline usage, which is 20%. Credits accumulate for as long as the instance runs at baseline usage. Moreover, an instance’s size affects how many CPU credits it receives. The larger the instance, the more CPU credits it accumulates over time.

What Are Amazon EC2 M5 Instances?

M5 instances are a type of General Purpose Amazon EC2 instance type to support medium and some high-performance applications. M5 instances are designed to provide a balanced combination of CPU, memory, and networking capabilities like other general purpose instances on AWS.

This focus on balance makes the instance type suitable for a wide variety of workloads. Yet M5 instances deliver more power than most general purpose instances, such as T3 instances.

With more power, comes the ability to handle medium-sized databases, cluster computing, and faster data processing.

The four types of M5 instances are M5, M5n, M5zn, and M5a — each with a unique contribution you’ll want to consider closely.

How Do T3 And M5 Instances Compare?

Here is a side-by-side comparison table showing the key differences between M5 vs R5 instances on AWS.

	T3 Instances	M5 Instances
Supported processors	3.1 GHz (Skylake 8175M or Cascade Lake 8259CL) Intel Xeon Scalable processorsAMD EPYC 7571 for T3a instances	3.1 GHz (Skylake 8175M or Cascade Lake 8259CL) Intel Xeon Scalable processorsAMD EPYC 7571 for M5a instances
Number of Cores	1 to 8 cores per instance	1 to 48 cores per instance
Maximum Memory	32 GiB	384 GiB
Maximum Storage	8 TiB EBS optimized	3.6 TiB NVMe based SSDs terabytes
Instance sizes	t3.nano (2 vCPU and 0.5 GiB memory)t3.micro (2 vCPU and 1 GiB memory)t3.small (2 vCPU and 2 GiB memory)t3.medium (2 vCPU and 4 GiB memory)t3.large (2 vCPU and 8 GiB memory)t3.xlarge (4 vCPU and 16 GiB memory)t3.2xlarge (8 vCPU and 32 GiB memory)	m5.large (2 vCPU and 8 GiB memory)m5.xlarge (4 vCPU and 16 GiB memory)m5.2xlarge (8 vCPU and 32 GiB memory)m5.4xlarge (16 vCPU and 64 GiB memory)m5.8xlarge (32 vCPU and 128 GiB memory)m5.12xlarge (48 vCPU and 192 GiB memory)m5.16xlarge (64 vCPU and 256 GiB memory)m5.24xlarge (96 vCPU and 394 GiB memory)m5.metal (96 vCPU and 394 GiB memory)
Networking	Up to 5 Gbps	Up to 25 Gbps with Enhanced Networking
Best Use cases	Small and medium databases, web and app servers, virtual desktops	Cluster computing, medium-sized databases, data processing

When To Use T3 Instances

If your workload requires moderate CPU usage but sees temporary spikes, you’ll want to consider Amazon EC2 T3 instances. These use cases can also include running microservices, development environments, and repositories.

When To Use M5 Instances

You might consider Amazon EC2 M5 instances if you have workloads that need balanced compute, memory, and network resources, but could use more power than T3 instances can deliver. Think: running mid-size databases, real-time big data processing, as well as high performance web and application servers.

How To Choose The Best AWS Instance Type For Your Application

Here’s the deal. You can go the manual route. That’ll require you to experiment with multiple instance types and sizes to figure out which combination works best.

Or, you can use CloudZero Advisor to rightsize your environment for optimal price-performance. See how CloudZero Advisor works here.

After that, you can easily track, understand, and optimize your AWS costs. Most cost platforms are inexact, clunky, and manual. With CloudZero, you get immediately actionable cost intelligence, such as cost per individual customer, per team, per feature, per project, and more.

Don’t take our word for it; take a tour of CloudZero here.

Ready to experience the only cloud cost optimization platform that does all of this for yourself? Schedule a demo today.