Skip to content
AWS Graviton is a line of processors designed by Amazon Web Services specifically optimized for cloud workloads running in Amazon Elastic Compute Cloud (EC2) applications. Built around ARM architecture, these processors deliver enhanced performance and efficiency for a wide range of cloud workloads, including application servers, microservices, cluster computing, and machine learning.
Graviton processors are known for providing the best price-performance ratio compared to traditional x86 processors, and architecting your cloud workloads for them will generally result in reduced costs.
Why did AWS create Graviton?
The AWS Graviton processors stem from AWS’s acquisition of Annapurna Labs in 2015. This partnership initially focused on the AWS Nitro System, which integrated specialized hardware with a lightweight hypervisor to maximize hardware power for Amazon Elastic Compute Cloud (EC2) instances. Following on from Nitro, Annapurna Labs under AWS developed its own custom CPU, designed to efficiently handle scale-out workloads.
This led to the introduction of the AWS Graviton processors, which were first launched with the A1 instances in 2018. These processors are based on the ARM architecture, known for its energy efficiency and lower costs compared to traditional x86-64 processors.
Graviton processors are optimized for many general purpose use cases and are built to integrate well with AWS’s infrastructure.
What are the benefits of AWS Graviton?
Enhanced Performance:
Cost Efficiency:
Energy Efficiency:
Broad Ecosystem Support:
Security:
AWS Graviton processors incorporate multiple security features, such as always-on 256-bit DRAM encryption, which helps in protecting data in transit and at rest. Additionally, these processors are designed to minimize the surface area for security attacks, thereby enhancing overall instance security.
Why do engineers still have to know EC2 instance types?
Graviton vs Graviton2 vs Graviton3 vs Graviton4
Graviton
- Released in 2018
- 16 Cortex A72 cores, with ARMv8-A ISA including Neon, crc, crypto.
- vCPUs are physical cores in a single NUMA domain, running at 2.3 GHz
- Only the A1 EC2 instance contains the first version of Graviton
Graviton2
- Released in 2019
- 64 Neoverse N1 cores, with ARMv8.2-A ISA including 2×128 bit Neon, LSE, fp16, rcpc, dotprod, crypto.
- vCPUs are physical cores in a single NUMA domain, running at 2.5 GHz
- EC2 instances with Graviton2 CPU: M6g, M6gd, C6g, C6gd, C6gn, R6g, R6gd, T4g, X2gd, G5g, Im4gn, Is4gen, I4g
- Offer 40% better price performance over comparable x86/x64 instances
Graviton3
- Released in 2022
- 64 Neoverse V1 cores, with ARMv8.4-A ISA including 4×128 bit Neon, 2×256 bit SVE, LSE, rng, bf16, int8, crypto.
- vCPUs are physical cores in a single NUMA domain, running at 2.6 GHz.
- EC2 instances with Graviton3 CPU: C7g, M7g, R7g, with Graviton3E (a higher powered version of Graviton3) CPU: C7gn, HPC7g
- Offer up to 25% better general performance over Graviton2
Graviton4
- Announced at Amazon re:Invent 2023
- 96 Neoverse V2 cores, with ARMv9.0-A ISA
- EC2 instances with Graviton4 CPU: R8g
- OFfer up to 30% better compute performance than Graviton3
How does Graviton pricing work?
AWS Graviton pricing primarily follows the standard AWS pricing model for EC2 instances, which is based on several factors:
Instance Type: The cost varies depending on the instance family (like M6g, C6g, R6g) and the specific instance size within that family. Each family is optimized for different types of workloads, such as general-purpose, compute-optimized, etc. Other instances are optimized for memory-intensive workloads or storage-intensive workloads.
Usage Time: Amazon EC2 instances are typically billed by the second, with a minimum of 60 seconds. This means you pay only for the compute time you use.
Region: AWS Graviton prices can vary by AWS region. For example, instances in North American regions might be priced differently from those in Asia or Europe.
Purchasing Options:
On-Demand: Pay for what you use without any upfront cost.
Reserved Instances/Savings Plans: Commit to a specific instance type in a region for a term of one or three years, with the option to pay all, some, or none of the cost upfront. This commitment can significantly reduce hourly costs.
Spot Instances: Purchase unused Amazon EC2 capacity at potentially significant discounts compared to the On-Demand price. Spot instance pricing is variable, based on supply and demand, and instances can be interrupted by AWS with a two-minute notification.
Operating System: Pricing for AWS Graviton instances varies based on the choice of operating system due to licensing costs, especially for instances running for example Windows compared to those running Linux or other open-source operating systems.
AWS Graviton FAQ
Are there any downsides to using AWS Graviton based instances?
In general, there aren’t many downsides to using Graviton instances as long as your application architecture is compatible with it.
In some cases Graviton instances may face compatibility issues with software designed solely for x86 architectures, necessitating significant migration efforts for existing applications. Also, Graviton instances only support Linux operating systems (not Windows, etc.). There are a few other edge cases such as performance variations for specific workloads not optimized for ARM, and limited availability of some specialized tools and services in the ARM ecosystem.
What AWS services support Graviton?
AWS services that support Graviton processors include
Amazon RDS: Supports Graviton processors for databases like MySQL, PostgreSQL, MariaDB, and Amazon Aurora, providing better price performance.
Amazon ElastiCache: Utilizes Graviton processors for Redis and Memcached, enhancing performance and reducing costs.
Amazon EMR: Graviton support allows for more efficient processing of big data workloads.
AWS Lambda: Graviton support for serverless compute functions.
Amazon OpenSearch Service: Benefits from Graviton for search and analytics workloads.
Amazon EKS (Elastic Kubernetes Service) and ECS (Elastic Container Service): Support Graviton for running containerized applications.
Amazon Neptune: Utilizes Graviton for graph database services
AWS Fargate: Supports Graviton2 processors for serverless container execution
Why do engineers still need to manually choose the right EC2 instance types?
If you find that choosing between all the different EC2 instance types isn’t scalable, you’re not alone. Choosing EC2 instance types can be time-consuming due to the wide range of options, considerations, and detailed knowledge required about each option’s strengths and limitations.
Here, tools like Compute Copilot can be incredibly beneficial. As an intelligent workload provisioner, Compute Copilot manages your compute resources while optimizing cost savings through the use of cost-effective Spot instances. By automatically recommending the most suitable instance types based on specific workload requirements, Compute Copilot eliminates the need to manually navigate the extensive AWS architecture options.
Once you specify your requirements, you instantly receive the full list of resources optimally tailored to your workloads. This feature is particularly invaluable for maximizing your Spot stability and savings with ASG or EKS.
(Tip: Why is Spot diversity important? Access to the largest, most diverse range of appropriate instance types is highly advantageous, because you’re not limiting yourself to a specific type or size, which may be in high demand and thus more likely to experience price spikes or be interrupted. By providing additional capacity at a lower cost to cover spikes in demand, Spot can help you make fewer purchase commitments to retain operational flexibility — while still ensuring you get discounts on all of your compute).
You can manually choose from the list of workload-suitable families if desired, or select all eligible instance families for the broadest possible pool of cheap and reliable Spot instances that Copilot will automatically move your workloads onto.
About nOps
nOps manages $1.5 billion in AWS spend, and was recently named #1 in G2’s Cloud Cost Management category.
Book a demo to find out how Compute Copilot can help you automatically manage and optimize your workloads for optimal performance and cost-efficiency.
