Trace3 Cloud FinOps: Save Money in an Instance!

By Chelsea Robertson | Senior Cloud FinOps Consultant, Trace3 Cloud Solutions
& Amanda Wagner | Associate Cloud FinOps Consultant, Trace3 Cloud Solutions 

 

In one of our recently published blogs we shared our teams’ experiences at AWS Re:Invent and included our thoughts on Graviton. As discussed, moving an instance to Graviton has cost advantages but can often reveal compatibility issues with existing code. Graviton is not the only alternative to Intel. In 2018, Advanced Micro Devices threw their hat in the ring by launching their first AMD EPYC line of processors on AWS. Their instances offer competitive pricing while maintaining or boosting performance and usually requires no architecture changes.

Last month, we had the opportunity to meet with the AMD team to learn about their compute offerings and how they have expanded their footprint. We reviewed performance, security, and compatibilities. We know you’re wondering, “what’s Trace3’s take on all of this?”


Changing instance types within your environment may be daunting to some, but AMD makes it accessible with 14 unique EC2 instance types in 23 out of 30 AWS regions as well as 70 availability zones. A huge plus to AMD is that it runs on the AMD64 (x86_64) architecture. Unlike moving to a Graviton instance running on ARM architecture, you won’t be faced with potentially redesigning your workload to migrate from Intel EC2 to AMD EC2 instances. In addition, AMD instances are 10% less expensive than comparable instances in most cases, and up to 45% less expensive in India regions. All without impacting performance.

Picture1-4

(AMD’s Instance type options on EC2)

In January of 2023, AMD achieved High Performance Computing (HPC) competency status through the AWS competency program, a standard that continues to evolve as various capabilities advance. AMD demonstrated technical proficiency and proven customer success in HPC solutions to achieve this competency.

Picture2-2

(Overview of AMD EPYC on AWS)

AMD has prioritized being a leader in the security space too. All AMD EPYC processors feature a suite of security features called AMD Infinity Guard. Infinity Guard encrypts the main memory, as well as the virtual machine, helping to prevent certain cold-boot attacks and provides a layer of protection for data confidentiality. AMD reports that this full memory encryption is achieved with high-performance encryption engines integrated into the memory channels, without compromising on performance.

Picture3

(Security infographic of AMD Infinity Guard)

As for VM encryption, this chip is currently the only x86 processor with built-in Secure Encrypted Virtualization – Secure Nested Paging (SEV – SNP), which encrypts each VM with one of up to 509 unique encryption keys. It is important to note that no changes or refactoring is required to take advantage of SEV-SNP on AMD. So, for the most security conscious customer looking to protect data-in-use, AMD EPYC-powered instances are an excellent choice.

Bottom Line

Due to the lack of awareness of AMD instance availability, this option is often overlooked. Search for instance types with the “a” (E.g.- M6a) to see if a more cost-effective and secure option with the same or better performance is available for your workload.

 To learn more about Trace3 Cloud FinOps, click here.



Robertson, Chelsea
Chelsea Robertson is a Senior FinOps Consultant on the Trace3 Cloud FinOps team with a background in data analysis and development work. This experience combined with a degree in Library Science and Information Systems means she has a propensity to take a deep dive into data and enjoys researching topics in emerging spaces such as the intersection between GreenOps and FinOps. She is passionate about supporting organizations in their quest to meet cloud cost savings initiatives and being a part of an experienced team working as trusted advisors in the field of FinOps. On a personal note, Chelsea has two kids, two dogs, one snake, and is newly dabbling in pottery as a hobby.
 

Wagner, Amanda
Amanda Wagner is a Cloud FinOps Consultant at Trace3, who brings her energy and previous experiences as an educator to guide her customers on their cloud journey. Drawing from said educational expertise, Amanda is motivated by the mission to serve her customers and help them continue to build on their knowledge of cloud management. In her time with Trace3, she has earned her FinOps Certified Practitioner and Cloud Health AWS Administrator Certifications. Amanda lives in Tampa with her husband, Matthew and dog, Penny. She enjoys tending to her caterpillar habitat, where she raises caterpillars until they hatch into Monarch butterflies and can be released.
Back to Blog