Blog / AWS

T4g Instances: Optimized Performance and Cost Efficiency for Burstable Workloads

T4g instances are a cost-effective and flexible solution for running burstable general-purpose workloads on Amazon EC2.

 · 3 min read

Introduction:

T4g instances are a cost-effective and flexible solution for running burstable general-purpose workloads on Amazon EC2. Powered by AWS Graviton2 processors and built on the AWS Nitro System, T4g instances offer a range of benefits, including superior price performance, enhanced security, and extensive ecosystem support.

Benefits:

  1. T4g instances offer up to 40% better price performance than T3 instances, making them cost-efficient for various applications.
  2. They provide 40% higher peak performance, suitable for smooth migration of production workloads.
  3. T4g instances are flexible and can run burstable general-purpose workloads like microservices and databases.
  4. Developers can build Arm applications natively in the cloud and benefit from EC2’s scalability and reliability.
  5. Enhanced security with 256-bit DRAM encryption and faster per-core encryption performance through AWS Graviton2 processors and AWS Nitro System .
  6. The lightweight hypervisor maximizes resource efficiency, improving overall performance.
  7. T4g instances are supported by popular Linux operating systems and work seamlessly with many AWS services and Independent Software Vendors (ISVs).


Features:

  • Performance

        We compared the performance between these instance families by running a bunch of tests on the following instances:

        t3.xlarge

       t3a.xlarge

       t4g.xlarge

Instance TypeNumber of coresMemory (GB)
t3.xlarge416
t3a.xlarge416
t4g.xlarge416

1. CPU

We ran the CPU benchmark using the following command:  sysbench cpu –threads=4 run. We will be comparing the total time as well as the total number of events for the benchmark from start to end.

Instance TypeTotal TimeTotal Number of Events
t3.xlarge10.0008s32591
t3a.xlarge10.0010s26621
t4g.xlarge10.0004s113588

2.Memory

We ran the CPU benchmark using the following command: sysbench memory –threads=4 run. We will be comparing the Operations performed for the different instance types.

In this benchmark, t3.xlarge performed the best whereas t4g.xlarge performed the worst.

Instance TypeTotal OperationsTotal Number of transferred in (MiB/sec)
t3.xlarge93129278 (9311302.73 per second)90946.56 MiB (9093.07 MiB/sec)
t3a.xlarge85354682 (8533783.52 per second)83354.18 MiB (8333.77 MiB/sec)
t4g.xlarge61109295 (6110003.56 per second)59677.05 MiB (5966.80 MiB/sec)

3. Disk performance

We also ran simple read & write benchmarks to test the performance of the EBS volumes on each of these instance types.

The T3a instances performed the best on the disk performance benchmarks.


  1. T3.xlarge

  Write : 1677721600 bytes (1.7 GB, 1.6 GiB) copied, 73.5538 s, 22.8 MB/s

  Read : 1677721600 bytes (1.7 GB, 1.6 GiB) copied, 44.5627 s, 37.6 MB/s


  1. T3a.xlarge

  Write  :  1677721600 bytes (1.7 GB, 1.6 GiB) copied, 66.9927 s, 25.0 MB/s

  Read  :  1677721600 bytes (1.7 GB, 1.6 GiB) copied, 39.1757 s, 42.8 MB/s


  1. T4g.xlarge

  Write : 1677721600 bytes (1.7 GB, 1.6 GiB) copied, 73.5538 s, 22.8 MB/s

  Read : 1677721600 bytes (1.7 GB, 1.6 GiB) copied, 65.7547 s, 25.5 MB/s


Pricing

We compared the prices between these instance families by looking at the price per hour for each instance type.

T4g instances are the cheapest amongst the three instance families, whereas T3 instances are the most expensive.

Price/core for Linux on-demand instances:

Instance TypePrice/hourPrice/monthly
t3.xlarge$0.1664$130.8160 monthly
t3a.xlarge$0.1504$71.9780 monthly
t4g.xlarge$0.1344$65.4080 monthly
  1. Unlimited and Standard Modes: T4g instances start in unlimited mode, offering sustained high CPU performance over any desired time frame. The hourly instance price covers interim spikes in usage. In standard mode, instances can burst until credits are depleted.


  1. Powered by AWS Graviton2 Processors: AWS Graviton2 processors, based on 64-bit Arm Neoverse cores, deliver 7x higher performance, 4x more compute cores, 2x larger caches, and 5x faster memory compared to the first-generation Graviton processors.


  1. Built on the AWS Nitro System: T4g instances leverage the AWS Nitro System, offloading virtualization functions to dedicated hardware and software. This results in high performance, availability, security, and reduced virtualization overhead.


Conclusion:

The new T4g instances provide better CPU performance, and they are cheaper than T3 and T3a instances. This should be the go-to instance type in this instance family.

T4g instances share the credits system, AWS Nitro System, and Burstable mode with T3 instances. However, they differentiate themselves with the latest Arm-based AWS Graviton2 processors. These custom-built processors offer up to 40% more performance than T3 instances, all at the same price.


A

No comments yet.

Add a comment
Ctrl+Enter to add comment