When engineering teams leverage cloud computing to run their workloads, it’s easy to overlook the research and validation required to understand which instance types might provide them the best price and performance fit.
Intel Tiber App-Level Optimization featured as a new Graviton Partner at AWS re:Invent’s Global Partner Summit Keynote
There are several reasons that contribute to this: the vast list of instance varieties, lack of cloud expertise, missing points of comparison data, and many others. Perhaps a more likely reason is that with cloud instances priced at a few dollars per hour, it’s simply easy for engineers to dismiss their relevance. However, with workloads often scaling with demand across multiple instances and environments, the cost of that instance type can range from impactful to budget-breaking.
Moreover, it’s easy for cloud engineers to think exclusively about virtual CPUs and memory while ignoring the hardware specs that support those virtual machines. With the introduction of AWS Graviton-powered instance types, Amazon Web Services brought a lot of attention to underlying price and performance gains between similar instance types when making a simple switch across different hardware computing architectures.
What is AWS Graviton?
Graviton is a type of computing processor designed and developed by Amazon Web Services, exclusively tailored to their global cloud infrastructure. Leveraging 64-bit Arm Neoverse cores, AWS Graviton processors are custom built to deliver the optimal price and performance ratio for workloads running in AWS EC2.
Graviton marked the introduction of Arm processors in the cloud world and extended EC2’s capabilities to support multiple compute architectures in addition to the existing last generation Intel and AMD processors (x86). Today, Graviton processors provide exclusive hardware and software security functions and are capable of operating at scale with always-on 256-bit DRAM encryption while delivering a 50% faster encryption performance per core compared with the previous generation.
Understanding the AWS Graviton Cost and Performance Gains
Graviton is much more than just another computing architecture for AWS EC2 with enhanced security capabilities. With the first generation of Graviton, AWS built a processor tailored to deliver compelling cost savings to customers using general-purpose instance types for workloads such as web servers, microservices and other types of applications requiring a small CPU and memory footprint.
However, with the second generation of Graviton, AWS significantly expanded its features to cover additional instance types such as compute-optimized (C6g, C6gd, C6gn), general-purpose burstable (T4g) and memory-optimized (R6g, R6gd, X2gd), including the variants with NVMe-based local SSD storage.
These new Graviton2 instances are capable of delivering up to 40% better price-performance compared to the equivalent x86-based instances. With the introduction of this wide range of instance types, Graviton is now capable of catering to different workload requirements such as high-performance computing, in-memory caching databases, video encoding and machine learning inference.
Graviton offers a great advantage to any AWS customer capable of running their workload using an Arm-based architecture. With a lower hourly price compared with equivalent instance types, Graviton can deliver 7x more performance, 4x more compute cores, 5x faster memory, and 2x larger cache.
What AWS Services Support Graviton?
Nowadays the Graviton technology has been expanded to several AWS managed services, across different domain areas. Great examples are container orchestration and registry services such as Amazon ECS, EKS and ECR, databases and big data platforms like RDS, Aurora, ElastiCache, OpenSearch and EMR, and services that provide continuous integration and deployment capabilities such as Amazon CodeBuild, CodeCommit, CodePipeline, and CodeDeploy.
Moreover, the Graviton ecosystem enabled Arm developers to build cloud-native applications without the need for typical error-prone and time-consuming emulation and cross-compilation for multi-processor architectures. This enables AWS technology partners such as Intel Tiber App-Level Optimization to offer innovative products and services that add value on top of AWS Graviton and deliver additional performance and cost optimization benefits for AWS customers.
How AWS Graviton and Intel Tiber App-Level Optimization Deliver Immediate Customer Value
As a leading platform for real-time continuous performance optimization in compute workloads and also as an AWS partner, Intel Tiber App-Level Optimization is in pole position to provide organizations the benefits of AWS Graviton and to supplement them with its own technology.
Environmentally sustainable computing is an important driver for organizations who seek computing optimization technologies, and Graviton is a key part of AWS strategy towards cloud efficiency. Intel Tiber App-Level Optimization gives customers a software boost on top of optimized AWS Graviton hardware. Together, these two technologies enable customers to deliver more value, with less computing costs and with far greater power efficiency.
The Intel Tiber App-Level Optimization algorithm automatically learns the resource usage patterns and data flows of applications. It then identifies and addresses contented resources, bottlenecks and prioritization opportunities out of the box, without any need for code changes, which in turn enables the handling of a to compute workload with 60% less servers while improving performance by 40%. This can have a massive performance and cost impact when customers are operating microservice-enabled applications on elastic infrastructure.
Intel Tiber App-Level Optimization is workload agnostic and runs seamlessly with multiple processor architectures. In addition, customers can enjoy Intel Tiber App-Level Optimization’s cost and performance optimizations regardless of their infrastructure, with out-of-the-box support for both cloud (AWS, Azure, GCP) and on-premises solutions, and special features tailored to Kubernetes, containers and big data workloads.
The ability to further optimize the costs and performance of workloads that are already running on top of AWS Graviton might sound too good to be true, but the real-world results speak by themselves — SundaySky, which is transforming the customer experience of the world’s most demanding brands, was able to achieve a cost reduction of 15% and a 20% CPU utilization reduction in only one week without additional R&D efforts.
SundaySky only needed to deploy the Intel Tiber App-Level Optimization agents in their infrastructure (using a single command-line operation) to immediately achieve computing efficiencies and cost savings.
Conclusion
With Intel Tiber App-Level Optimization as an official AWS Graviton partner, customers can expect the best of both worlds when it comes to optimizing their costs and processing performance in AWS.
It is worth noting that Graviton was only launched in late 2018 with AWS EC2 instance support, yet since then it has been rapidly expanding across several other AWS managed services. The recent release of Graviton2 support for AWS Lambda, which promises up to 34% better price performance, comes to validate the positive impact this technology brings to different types of services and the potential for future expansion across the whole AWS managed service offering.