Marcos Ortiz

AWS Graviton Weekly # 24: Week from February 10th, 2023 to February 17th, 2023

publishedabout 1 month ago
9 min read

You're receiving this because you subscribed here OR here

This email may contain affiliate links. I receive a small commission for recommending products I use & love at no extra cost to you.

[Read the browser version right here]

Brought to you in partnership with Deel

Issue # 24: February 10th, 2023 to February 17th, 2023

Hey Reader

Welcome to Issue # 24 of AWS Graviton Weekly, which will be focused on sharing everything that happened in the past week related to AWS Silicon: from February 10th, 2023 to February 17th, 2023.

Today (February 16th, 2023) I had an incredible meeting with a couple of amazing entrepreneurs in the tech space from LATAM.

I can't share more about it, but I'm particularly excited for having the chance to chat with them about the future of software development driven by chips, chips design and of course: the amazing growth and influence that Arm is setting right now in the industry.

Stay tuned for more of this because we are cooking something very cool and you will one of the first ones to know about it.

Anyways: Enjoy the content of this week inside the Graviton universe.

Brought to you by Deel

Through Deel’s platform, you can hire employees and contractors in 150+ countries.

You can easily create contracts that are compliant with the local labor laws and streamline payroll for your entire remote team in one single click.

Run your global payroll with Deel


Introducing Amazon EC2 M7g and R7g Instances

AWS announces the general availability of Amazon Elastic Compute Cloud (Amazon EC2) M7g instances and Amazon EC2 R7g instances.
These instances are powered by AWS Graviton3 processors and deliver up to 25% better performance over Graviton2-based instances.
Compared to AWS Graviton2 processors, they offer up to 2x better floating-point performance, up to 2x faster crypto performance, and up to 3x better machine learning (ML) performance, including support for bfloat16.
These Graviton3-based instances feature the latest DDR5 memory, which provides 50% more memory bandwidth than DDR4, to enable high-speed access to data in memory.

Learn more here:

AWS announces AWS Modular Data Center (MDC) for U.S. Department of Defense customers

Today, AWS announces the availability of AWS Modular Data Center (MDC) to the U.S. Department of Defense (DoD). AWS MDC makes it easy for DoD agencies to deploy modular data centers managed by AWS in infrastructure-limited locations. AWS MDC is available as a self-contained modular data center unit: an environmentally controlled physical enclosure that can host racks of AWS Outposts or AWS Snow Family devices. AWS MDC can scale further through the deployment of additional units. Customers can proactively monitor and manage their modular data centers using the management system available with every unit. Customers also have the option to use satellite communications for network connectivity.

Learn more here:

Amazon Managed Grafana now supports network access control

Amazon Managed Grafana now supports inbound network access control that helps you to restrict user access to your Grafana workspaces.
Amazon Managed Grafana is a fully managed service for Grafana, a popular open-source analytics platform that enables you to query, visualize, and alert on your metrics, logs, and traces.
With this launch, you have granular security controls over the rollout of Grafana workspaces by defining customer-managed prefix lists and VPC endpoints to help you restrict the inbound network traffic that can reach your Grafana workspaces.

Learn more here:

Amazon EC2 Dedicated Hosts now support automated maintenance on rare degradation

You now have the automated host maintenance feature available for your Amazon EC2 Dedicated Hosts. With automated host maintenance, in the rare event of degradation of a dedicated host, AWS will automatically reboot the EC2 instances running on it onto a newly allocated dedicated host during a scheduled maintenance event, to reduce your application’s downtime and offload undifferentiated heavy-lifting of host maintenance.
AWS regularly monitors the health of your hosts and the instances running on them. With automated host maintenance, in the rare event of a degradation or for planned EC2 maintenance, AWS will notify you, allocate a new host, and schedule a maintenance event for two weeks later to give you time to prepare. During the maintenance event, AWS will reboot your instances on the replacement host and notify you if an instance cannot be moved. Along with reviewing the event details on the EC2 Events page or the AWS Health Dashboard, you can change the order of instance reboot or reschedule the event.

Learn more here:

Articles and Tutorials

New Graviton3-Based General Purpose (m7g) and Memory-Optimized (r7g) Amazon EC2 Instances, by Jeff Barr (Chief Evangelist at AWS)

Today I am happy to tell you about the newest Amazon EC2 instance types, the M7g and the R7g. Both types are powered by the latest generation AWS Graviton3 processors, and are designed to deliver up to 25% better performance than the equivalent sixth-generation (M6g and R6g) instances, making them the best performers in EC2.
The M7g instances are for general purpose workloads such as application servers, microservices, gaming servers, mid-sized data stores, and caching fleets. The R7g instances are a great fit for memory-intensive workloads such as open-source databases, in-memory caches, and real-time big data analytics.

Learn more here:

Scaling Large Language Model (LLM) training with Amazon EC2 Trn1 UltraClusters by KC Tung (Cloud Solution Architect at AWS and author of two Tensorflow books), Jeff Huynh (Principal Engineer in AWS Annapurna Labs) and Shruti Koparkar (Senior Product Marketing Manager at AWS)

In this post, we discussed how scaling your training job across an Trn1-UltraCluster, which is powered by Trainium accelerators in AWS, reduces the time to train a model. We also provided a link to the Neuron samples repository, which contains instructions on how to deploy a distributed training job for a BERT-Large model.
Trn1-UltraCluster runs distributed training workloads to train ultra-large deep learning models at scale. A distributed training setup results in much faster model convergence as compared to training on a single Trn1 instance.

Learn more here:

Achieve up to 27% better price-performance for Spark workloads with AWS Graviton2 on Amazon EMR Serverless, with Karthik Prabhakar (Senior Big Data Solutions Architect for Amazon EMR at AWS) and Nithish Kumar Murcherla (Senior Systems Development Engineer on the Amazon EMR Serverless team)

As demonstrated in this post, Graviton2 on EMR Serverless applications consistently yielded better performance for Spark workloads. Graviton2 is available in all Regions where EMR Serverless is available.
To see a list of Regions where EMR Serverless is available, see the EMR Serverless FAQs

Learn more here:

Build and Deploy Multi-Architecture Graviton Container Images for Spring Boot Microservices, by Wenqi Glantz (Software Architect at ArisGlobal)

A step-by-step guide on graviton image building and deployment on ECS Fargate

Odigos Version v0.1.4, by the keyval team

Today we are excited to announce version v0.1.4 of Odigos, an open-source project that automatically generates distributed traces for all applications within minutes without code changes.
This version focuses on ARM Support, new destinations and stability improvements.
ARM Support
ARM processors are becoming more and more popular and are now supported by Odigos. You can now run Odigos on the latest Macbooks, AWS Graviton, and other ARM based processors. This is a big step forward for Odigos that now supports applications written in Java, Go, Python, NodeJS and .NET on both x86-64 and ARM.

Learn more about Odigos here:

Ahana Cloud for Presto

Ahana Cloud for Presto supports AWS Graviton2-based Amazon Elastic Compute Cloud (Amazon EC2) instances for Presto coordinator and worker nodes in Ahana-managed Presto clusters. Graviton2 processors are designed by AWS to provide high price-performance for cloud workloads running in Amazon EC2.

Options to install R on a Raspberry Pi and other ARM systems, by Andres Castro Socolich

The options described here are mostly applicable to the Raspberry Pi family of Single Board Computers but since most options also work on Ubuntu distributions they can also be applicable to other Linux ARM systems like other SBCs or even “full-size” ARM computers like Ampere Altra or AWS Graviton.

Real-time Ubuntu is now generally available, by Canonical

“Accessing an industry-leading Linux OS on AWS, with the power of AWS Graviton processors, will help further hardware parity in automotive development”, said Stefano Marzani, Worldwide Technical Lead, Software-Defined Vehicles at AWS. “By working with Canonical, and providing access to real-time Ubuntu in the cloud, AWS is helping to empower developers to start implementing systems requiring deterministic behaviour- like automotive and industrial workloads – ahead of hardware availability and in parallel with hardware provisioning”.

Softserve Joins Aws Graviton Service Delivery Program

Softserve continues to expand its Amazon Web Services (AWS) capabilities and has announced that it earned the AWS Service Delivery designation for AWS Graviton, recognizing that SoftServe provides deep technical knowledge, experience, and proven success in delivering AWS Graviton-based Amazon Elastic Compute Cloud (Amazon EC2) instances. Achieving the AWS Service Delivery designation for AWS Graviton differentiates SoftServe as an AWS Partner Network (APN) member that is helping customers accelerate and scale their adoption of AWS Graviton so they can realize the price performance benefits sooner across more workloads.

AWS Re:Invent Session Recap: Observability, Security, and Feature Management, by LaunchDarkly

In November 2022, AWS held its annual Re:Invent conference, and, as usual, LaunchDarkly was there (Alex Hardman provided a quick recap of our involvement here.) One particular session involved Andrew Krug, Head of Security Advocacy at Datadog.
With some help from the aforementioned Alex, Andrew led a session on observability, security, and feature management. The session covered how threats have evolved, new detection methods, and how feature flags can play a crucial role in both detecting and responding to attacks in highly innovative (and sometimes entertaining) ways.
We encourage you to watch the event in its entirety when you have the time, but for now, here are some of the main highlights from Andrew and Alex’s session.

Slides, Videos and Audio

[WEBINAR] Getting started with your Cost Financial Management journey: AWS tooling demo, with Stephanie Gooch (Sr. Commercial Architect in OPTICS at AWS & Host of The Keys to AWS Optimization on Twitch) and Connor Murphy (Customer Optimization & Enablement at AWS)

In this webinar, we'll take you through each of the four AWS CFM principles: See, Save, Plan, and Run.
We’ll walk you through an example optimization (using Graviton) to shed light on the framework for how you should think about each CFM pillar and layer tooling into your strategy.

[VIDEO] Cost Optimization for OpenSearch Workloads, by Gene Alpert (AWS Data & Analytics Specialist at AWS)

Learning Objectives:
- Objective 1: Learn how storage tiering options help you control cost
- Objective 2: See how OpenSearch Serverless can remove overprovisioning costs
-Objective 3: Review how Graviton can help reduce overall costs

Download the presentation from here:

Direct cost savings with Amazon OpenSearch Service

[VIDEO] ElasticON EMEA: Cost saving and benefits with out-of-the-box, real-time full-stack profiling, with Sean Heelan (Principal Software Engineer at Elastic) and Israel Ogbole (Principal Product Manager at Elastic)

Today's complex application environment has hidden problems that slow performance for users, wastefully increase infrastructure costs and needlessly generating emissions. Widely deployed support for kernel technology eBPF (Extended Berkeley Packet Filter) now provides a hyper-efficient way to understand how compute cycles are being spent in a system, as well as the application bottlenecks. In this session, we'll review technology foundations and share how Elastic users are looking across their entire deployment to eliminate waste in their production systems.


And if you are a company looking for new members for your time, you can get access to our Talent Collective here.

There are 24 active candidates ready for interviews:


Join us to learn about building a disaster recovery solution using AWS Elastic Disaster Recovery (AWS DRS). Learn how AWS DRS helps customers minimize downtime and data loss with fast, reliable recovery of on-premises and cloud-based applications using affordable storage, minimal compute, and point-in-time recovery.
The discussion will offer 100–200 level solution and the value of building a disaster recovery solution. At the end, we will play the Kahoot game, and the winners will be awarded an Uber Eats gift card.

Quote of the week

“We were able to get 50% cost savings compared to our on-premises systems by running our data intensive satellite imagery processing application on AWS Graviton2-based R6g instances due to their superior performance.
Adopting AWS Graviton2 required no code changes and the price performance benefits will now enable regular updates and quick turnaround on future product enhancements.”

Peter Scarth (Principal Research Scientist at Cibo Labs)

Source: Amazon