AWS Graviton is a family of processors designed to deliver the best price performance for your cloud workloads running in Amazon Elastic Compute Cloud (Amazon EC2). These instances deliver strong performance at a lower cost compared to traditional x86-based options. This post explores how to leverage Graviton to optimize costs and enhance the performance of your AWS-managed services, with a specific focus on database migration strategies and containerized workloads.
Graviton processors offer several advantages:
Cost-effectiveness: Graviton instances consistently provide cost savings compared to x86 counterparts. This translates directly to lower costs for running your applications.
Performance: Graviton processors often deliver equal or better performance than x86 processors at a lower cost. This means you can get more compute power for your money. Graviton incorporates hardware accelerators for common cloud workloads like compression/decompression and encryption/decryption. By handling these tasks in fixed-function hardware blocks, Graviton offloads substantial work from the main CPU cores.
Efficiency: Graviton processors are designed to be more energy-efficient, consuming up to 60% less energy than comparable x86 instances. AWS Graviton chips are fabricated on a leading edge 7nm manufacturing process, making them extremely dense and power efficient. The Graviton design emphasizes high core counts and memory bandwidth to excel at scale-out workloads. This leads to lower energy bills and a smaller carbon footprint.
Seamless Integration: AWS offers a growing number of managed services optimized for Graviton, including RDS, ElastiCache, EMR, and more. This means you can easily leverage the cost benefits of Graviton across various services without significant migration efforts.
Amazon RDS and Amazon Aurora
Run your MySQL, PostgreSQL, or Aurora databases with reduced costs and improved performance by switching to Graviton-based instances. This transition can be done seamlessly with minimal downtime by leveraging read replicas.
AWS reports up to 35% performance improvement and up to 52% price-performance improvement for RDS open-source databases.
Steps to execute for seamless migration to Graviton with Read Replicas
Create a Graviton-based Read Replica:
In your RDS console, create a read replica of your existing x86-based database instance. Choose a Graviton instance class (e.g., db.m6g or db.r6g) for the read replica.
This process will replicate data from your primary instance to the Graviton read replica with minimal impact on your application.
Test and Monitor:
Once the read replica is available, direct some read traffic to it and monitor its performance. Ensure your application functions correctly with the Graviton instance.
This step helps you identify any potential compatibility issues and ensures a smooth transition.
Failover / Promote Replica to Primary:
When you're confident in the Graviton read replica's performance, execute a Failover for Aurora, and Promote the read replica for RDS instances so that it becomes the new primary.
For Aurora AWS will handle the failover process, redirecting traffic from the old primary (x86) to the new primary (Graviton).
For RDS instances before you promote your read replica, stop transactions on the source and wait for the replica lag to be 0. The promotion process takes a few minutes. After promotion, you can't restart replication using the promoted database as a target, as the read replica will be promoted as a standalone RDS instance.
Shut Down the x86 Instance:
Once the Graviton instance is serving as the primary and you've verified everything is working correctly, shut down the old x86-based instance.
This step completes the migration and starts generating cost savings.
Amazon ElastiCache
Improve the performance and reduce the costs of your Valkey, Memcached or Redis deployments by simply switching to Graviton-based Amazon ElastiCache. AWS reports migration to Graviton-based instances provide up to 40% better price performance while using up to 60% less energy than comparable EC2 instances.
OpenSearch Service
Graviton-based OpenSearch Service offers the same features and functionality as the x86-based version but with the added benefits of Graviton's price performance. From a performance perspective, these instances improve over previous generation instances, such as a 38% improvement in indexing throughput, a 50% reduction in indexing latency, and a 40% improvement in query performance, depending on the instance family and size. This is when compared to the x86-based alternatives.
By following AWS's migration guidelines, migrating to Graviton-based OpenSearch can be achieved with minimal disruption.
AWS Lambda
Reduce the cost of running your serverless applications and experience potentially faster execution times with Graviton-powered Lambda functions.
Memory (GB) | Graviton Request Charge (per 1M requests) | x86 Request Charge (per 1M requests) | Direct Savings |
1 | $0.33 | $0.37 | 10.8% |
2 | $0.47 | $0.53 | 11.3% |
Note: Pricing quotes here are for the us-east-1 region and may vary depending on the configuration.
Amazon Elastic Container Service (Amazon ECS) with Fargate
Deploy your containerized applications on Graviton-based ECS clusters for cost efficiency and enhanced performance. With ECS, you can leverage Fargate, a serverless compute engine for containers, to further simplify operations and reduce costs.
By choosing Graviton-based ECS Fargate tasks, you can benefit from cost savings and improved performance.
Comparing Graviton and x86 Pricing for ECS Fargate
vCPU | Memory (GB) | Graviton | x86 | Savings |
1 | 2 | $0.039 per hour | $0.049 per hour | 20% |
2 | 4 | $0.079 per hour | $0.099 per hour | 20% |
Note: Pricing quotes here are for the us-east-1 region and may vary depending on the configuration.
Amazon EC2
For workloads that require EC2 instances, consider using Graviton-based instances (e.g., m6g, c6g, r6g) to achieve cost savings. You can launch Graviton instances directly from the EC2 console and migrate your applications to them.
Comparing Graviton and x86 Pricing for EC2
Here's a comparison of on-demand pricing for popular EC2 instance types:
Instance Type | Graviton | x86 | Savings |
m6g.large m6i.large | $0.077 per Hour | $0.096 per Hour | 20% |
c6g.large c6i.large | $0.068 per Hour | $0.085 per Hour | 20% |
r6g.larger6i.large | $0.101 per Hour | $0.126 per Hour | 20% |
Note: Pricing quotes here are for the us-east-1 region and may vary depending on the configuration.
Real-World Success Stories
Case-studies from companies like CircleCI, Honeycomb, Confluent, CrowdStrike, Datadog and many more have achieved significant cost and performance improvements by migrating to Graviton.
Results and Benefits
Graviton offers a compelling opportunity to optimize AWS cloud costs. Start exploring Graviton and experience its benefits. You can achieve significant cost savings and improve performance by leveraging Graviton-based instances. Learn more about how we can help you save on your cloud cost at https://www.severalclouds.com/aws-cost-optimization