Today, Amazon Elastic Container Registry (Amazon ECR) announces registry policy v2 which now supports managing IAM permissions for all ECR API actions. This new registry policy makes it easier for customers to control usage of ECR capabilities within their accounts.

ECR registry policy allows customers to control usage of ECR private registries by granting permissions to perform registry-level actions to an AWS IAM principal. Registry policy version 1 (v1), only supported three actions: ReplicateImage, BatchImportUpstreamImage, and CreateRepository. Now, the new registry policy version 2 (v2) supports every ECR action. Using registry policy v2 makes it easier for customers to control permissions across all repositories in an ECR registry, allowing them to improve their security posture and save time versus configuring permissions individually across multiple repositories.

ECR registry policy v2 is now available for all ECR registries in all AWS commercial regions. You can migrate from registry policy v1 to v2 using the ECR management console or with the new ECR put-account-setting API. New ECR accounts will automatically use registry policy v2. To learn more about ECR’s registry policy and permissions, see our documentation.
 

 

​Today, Amazon Elastic Container Registry (Amazon ECR) announces registry policy v2 which now supports managing IAM permissions for all ECR API actions. This new registry policy makes it easier for customers to control usage of ECR capabilities within their accounts. ECR registry policy allows customers to control usage of ECR private registries by granting permissions to perform registry-level actions to an AWS IAM principal. Registry policy version 1 (v1), only supported three actions: ReplicateImage, BatchImportUpstreamImage, and CreateRepository. Now, the new registry policy version 2 (v2) supports every ECR action. Using registry policy v2 makes it easier for customers to control permissions across all repositories in an ECR registry, allowing them to improve their security posture and save time versus configuring permissions individually across multiple repositories. ECR registry policy v2 is now available for all ECR registries in all AWS commercial regions. You can migrate from registry policy v1 to v2 using the ECR management console or with the new ECR put-account-setting API. New ECR accounts will automatically use registry policy v2. To learn more about ECR’s registry policy and permissions, see our documentation.    

AWS customers can now access the Llama 3.3 70B model from Meta through Amazon SageMaker JumpStart. The Llama 3.3 70B model balances high performance with computational efficiency. It also delivers output quality comparable to larger Llama versions while requiring significantly fewer resources, making it an excellent choice for cost-effective AI deployments.

Llama 3.3 70B features an enhanced attention mechanism that substantially reduces inference costs. Trained on approximately 15 trillion tokens, including web-sourced content and synthetic examples, the model underwent extensive supervised fine-tuning and Reinforcement Learning from Human Feedback (RLHF). This approach aligns outputs more closely with human preferences while maintaining high performance standards. According to Meta, this efficiency gain translates to nearly five times more cost-effective inference operations, making it an attractive option for production deployments.

Customers can deploy Llama 3.3 70B through the SageMaker JumpStart user interface or programmatically using the SageMaker Python SDK. SageMaker AI’s advanced inference capabilities help optimize both performance and cost efficiency for your deployments, allowing you to take full advantage of Llama 3.3 70B’s inherent efficiency while benefiting from a streamlined deployment process.

The Llama 3.3 70B model is available in all AWS Regions where Amazon SageMaker AI is available. To learn more about deploying Llama 3.3 70B on Amazon SageMaker JumpStart, see the documentation or read the blog.
 

 

​AWS customers can now access the Llama 3.3 70B model from Meta through Amazon SageMaker JumpStart. The Llama 3.3 70B model balances high performance with computational efficiency. It also delivers output quality comparable to larger Llama versions while requiring significantly fewer resources, making it an excellent choice for cost-effective AI deployments. Llama 3.3 70B features an enhanced attention mechanism that substantially reduces inference costs. Trained on approximately 15 trillion tokens, including web-sourced content and synthetic examples, the model underwent extensive supervised fine-tuning and Reinforcement Learning from Human Feedback (RLHF). This approach aligns outputs more closely with human preferences while maintaining high performance standards. According to Meta, this efficiency gain translates to nearly five times more cost-effective inference operations, making it an attractive option for production deployments. Customers can deploy Llama 3.3 70B through the SageMaker JumpStart user interface or programmatically using the SageMaker Python SDK. SageMaker AI’s advanced inference capabilities help optimize both performance and cost efficiency for your deployments, allowing you to take full advantage of Llama 3.3 70B’s inherent efficiency while benefiting from a streamlined deployment process. The Llama 3.3 70B model is available in all AWS Regions where Amazon SageMaker AI is available. To learn more about deploying Llama 3.3 70B on Amazon SageMaker JumpStart, see the documentation or read the blog.    

Amazon Lightsail API endpoints now support the IPv6 protocol, allowing you to connect over IPv6. To use this new capability, point your applications to use Amazon Lightsail’s new dual-stack endpoints. When you make a request to a dual-stack endpoint, the endpoint URL resolves to an IPv6 or an IPv4 address, depending on the protocol used by your network and client.

Public IPv4 addresses are being exhausted with the growth of the Internet. Earlier this year, Lightsail launched IPv6-only instance plans to support IPv6 adoption. With the introduction of dual-stack Lightsail API endpoints, you can now make requests to the Lightsail API from your Lightsail IPv6-only instances, or any IPv6 client.

You can use this capability with the AWS Command Line Interface (CLI) and AWS SDKs in all AWS Regions supporting Lightsail. To learn more, please see AWS Service Endpoints documentation and Lightsail Service Endpoints documentation.

 

​Amazon Lightsail API endpoints now support the IPv6 protocol, allowing you to connect over IPv6. To use this new capability, point your applications to use Amazon Lightsail’s new dual-stack endpoints. When you make a request to a dual-stack endpoint, the endpoint URL resolves to an IPv6 or an IPv4 address, depending on the protocol used by your network and client. Public IPv4 addresses are being exhausted with the growth of the Internet. Earlier this year, Lightsail launched IPv6-only instance plans to support IPv6 adoption. With the introduction of dual-stack Lightsail API endpoints, you can now make requests to the Lightsail API from your Lightsail IPv6-only instances, or any IPv6 client. You can use this capability with the AWS Command Line Interface (CLI) and AWS SDKs in all AWS Regions supporting Lightsail. To learn more, please see AWS Service Endpoints documentation and Lightsail Service Endpoints documentation.  

Amazon Bedrock Agents, Flows, and Knowledge Bases now offers support for the recently announced, in-preview, latency-optimized models via the SDK. This enhancement brings faster response times and improved responsiveness to AI applications built with Amazon Bedrock Tooling. Currently, this optimization is available for Anthropic’s Claude 3.5 Haiku model and Meta’s Llama 3.1 405B and 70B models, delivering reduced latency compared to standard models without compromising accuracy.

This update is particularly beneficial for customers developing latency-sensitive applications such as real-time customer service chatbots and interactive coding assistants. By leveraging purpose-built AI chips like AWS Trainium2 and advanced software optimizations in Amazon Bedrock, customers can now access more options to optimize their inference for specific use cases. Importantly, these capabilities can be integrated immediately into existing applications without additional setup or model fine-tuning, resulting in enhanced performance and faster response times.

The latency-optimized inference support for Amazon Bedrock Agents, Flows, and Knowledge Bases is available in the US East (Ohio) Region via cross-region inference. Customers can access these new capabilities through the Amazon Bedrock SDK via a runtime configuration, enabling them to programmatically incorporate these optimized models into their workflows and applications.

To learn more about Amazon Bedrock and its capabilities, including this new latency-optimized inference support, visit the Amazon Bedrock product page, pricing page, and documentation.
 

 

​Amazon Bedrock Agents, Flows, and Knowledge Bases now offers support for the recently announced, in-preview, latency-optimized models via the SDK. This enhancement brings faster response times and improved responsiveness to AI applications built with Amazon Bedrock Tooling. Currently, this optimization is available for Anthropic’s Claude 3.5 Haiku model and Meta’s Llama 3.1 405B and 70B models, delivering reduced latency compared to standard models without compromising accuracy. This update is particularly beneficial for customers developing latency-sensitive applications such as real-time customer service chatbots and interactive coding assistants. By leveraging purpose-built AI chips like AWS Trainium2 and advanced software optimizations in Amazon Bedrock, customers can now access more options to optimize their inference for specific use cases. Importantly, these capabilities can be integrated immediately into existing applications without additional setup or model fine-tuning, resulting in enhanced performance and faster response times. The latency-optimized inference support for Amazon Bedrock Agents, Flows, and Knowledge Bases is available in the US East (Ohio) Region via cross-region inference. Customers can access these new capabilities through the Amazon Bedrock SDK via a runtime configuration, enabling them to programmatically incorporate these optimized models into their workflows and applications. To learn more about Amazon Bedrock and its capabilities, including this new latency-optimized inference support, visit the Amazon Bedrock product page, pricing page, and documentation.    

You can now create Amazon Managed Streaming for Apache Kafka (Amazon MSK) provisioned clusters running on AWS Graviton3-based M7g instances or upgrade your existing x-86 based based M5 or T3 instances and replace them with AWS Graviton3-based M7G instances with a single click of a button in the Europe (Paris) AWS Region.

AWS Graviton3 processor based M7G instances on Amazon MSK provisioned clusters allows you to achieve up to 24% compute cost savings and up to 29% higher write and read throughput over comparable MSK clusters running on M5 instances. Additionally, these instances lower energy consumption by up to 60% than comparable instances, making your Kafka clusters more environmentally sustainable.

Please refer to our blog for more information on the price/ performance improvements of M7g instances and the Amazon MSK pricing page for information on pricing. To get started, you can create new clusters using M7G instances or update your existing clusters to M7G brokers using the AWS Management Console, and read our developer guide for more information.
 

 

​You can now create Amazon Managed Streaming for Apache Kafka (Amazon MSK) provisioned clusters running on AWS Graviton3-based M7g instances or upgrade your existing x-86 based based M5 or T3 instances and replace them with AWS Graviton3-based M7G instances with a single click of a button in the Europe (Paris) AWS Region. AWS Graviton3 processor based M7G instances on Amazon MSK provisioned clusters allows you to achieve up to 24% compute cost savings and up to 29% higher write and read throughput over comparable MSK clusters running on M5 instances. Additionally, these instances lower energy consumption by up to 60% than comparable instances, making your Kafka clusters more environmentally sustainable. Please refer to our blog for more information on the price/ performance improvements of M7g instances and the Amazon MSK pricing page for information on pricing. To get started, you can create new clusters using M7G instances or update your existing clusters to M7G brokers using the AWS Management Console, and read our developer guide for more information.    

SES Mail Manager now offers end to end logging for ingress endpoints and rules engine actions. Customers have the ability to configure a wide range of monitoring options to the three standard logging destinations: CloudWatch, S3, and Firehose.

Mail Manager by design interoperates with external systems on both incoming and outgoing email flows. Customers expect to use it to gain visibility on email volumes as well as to troubleshoot configuration problems at every step in the email delivery paths. These new logging features enable both ad hoc investigations and automated alarming via standard CloudWatch tooling, ensuring that the overall integrity of Mail Manager configurations can be tracked programmatically alongside other customer infrastructure. The logging features also help troubleshooting in the event of a configuration change on a connected system, reducing the support burden and enabling more customer self-service solutions.

Mail Manager logging is available in every AWS Region where Mail Manager is launched. Customers can learn more about Mail Manager by clicking here.
 

 

​SES Mail Manager now offers end to end logging for ingress endpoints and rules engine actions. Customers have the ability to configure a wide range of monitoring options to the three standard logging destinations: CloudWatch, S3, and Firehose. Mail Manager by design interoperates with external systems on both incoming and outgoing email flows. Customers expect to use it to gain visibility on email volumes as well as to troubleshoot configuration problems at every step in the email delivery paths. These new logging features enable both ad hoc investigations and automated alarming via standard CloudWatch tooling, ensuring that the overall integrity of Mail Manager configurations can be tracked programmatically alongside other customer infrastructure. The logging features also help troubleshooting in the event of a configuration change on a connected system, reducing the support burden and enabling more customer self-service solutions. Mail Manager logging is available in every AWS Region where Mail Manager is launched. Customers can learn more about Mail Manager by clicking here.    

AWS CloudTrail introduces dual stack support for the CloudTrail API endpoints, enabling you to connect using Internet Protocol Version 6 (IPv6), Internet Protocol Version 4 (IPv4), or dual stack clients. Dual stack support is also available when you privately access the CloudTrail API endpoint from your Amazon Virtual Private Cloud (VPC) using AWS PrivateLink.

The urgency to transition to Internet Protocol version 6 (IPv6) is driven by the continued growth of internet, which is exhausting available Internet Protocol version 4 (IPv4) addresses. With simultaneous support for both IPv4 and IPv6 clients on CloudTrail endpoints, you are able to gradually transition from IPv4 to IPv6 based systems and applications, without needing to switch all over at once. This enables you to meet IPv6 compliance requirements and removes the need for expensive networking equipment to handle the address translation between IPv4 and IPv6

To learn more on best practices for configuring IPv6 in your environment, visit the whitepaper on IPv6 in AWS. Support for IPv6 on AWS CloudTrail is available in all commercial regions and the AWS GovCloud (US) Regions.

 

​AWS CloudTrail introduces dual stack support for the CloudTrail API endpoints, enabling you to connect using Internet Protocol Version 6 (IPv6), Internet Protocol Version 4 (IPv4), or dual stack clients. Dual stack support is also available when you privately access the CloudTrail API endpoint from your Amazon Virtual Private Cloud (VPC) using AWS PrivateLink. The urgency to transition to Internet Protocol version 6 (IPv6) is driven by the continued growth of internet, which is exhausting available Internet Protocol version 4 (IPv4) addresses. With simultaneous support for both IPv4 and IPv6 clients on CloudTrail endpoints, you are able to gradually transition from IPv4 to IPv6 based systems and applications, without needing to switch all over at once. This enables you to meet IPv6 compliance requirements and removes the need for expensive networking equipment to handle the address translation between IPv4 and IPv6 To learn more on best practices for configuring IPv6 in your environment, visit the whitepaper on IPv6 in AWS. Support for IPv6 on AWS CloudTrail is available in all commercial regions and the AWS GovCloud (US) Regions.  

AWS has launched new Guides on AWS Partner Central providing detailed resources with step-by-step guidance on AWS Partner Network and Marketplace topics. The new guides provide introductory and advanced resources for every stage of your AWS journey, giving you the guidance you need to move up and across the AWS Partner Profitability Framework. In addition, guides provide enablement materials for Managed Service Provider, Solution Provider and ISV Accelerate Program, guidance on using AWS Co-sell Experience (ACE) and AWS Partner Funding Portal (APFP), and information on key topics, such as Generative AI and migration and modernization. These improvements provide Partners with structured, actionable information and quick access to resources, making it easier to navigate their AWS partnership.

Partners can access guides in AWS Partner Central while completing assigned Tasks or by using Partner Assistant, the generative AI virtual assistant. From Task descriptions or Partner Assistant responses, partners can navigate directly to relevant guides for in-depth information and related resources. This integrated experience helps partners to find answers quickly, progress through their AWS Partner journey more efficiently and unlock benefits faster.

Guides are available to all partners. To get started, log into AWS Partner Central, navigate to the «Resources» menu, and select «Guides». Learn more about becoming an AWS Partner.
 

 

​AWS has launched new Guides on AWS Partner Central providing detailed resources with step-by-step guidance on AWS Partner Network and Marketplace topics. The new guides provide introductory and advanced resources for every stage of your AWS journey, giving you the guidance you need to move up and across the AWS Partner Profitability Framework. In addition, guides provide enablement materials for Managed Service Provider, Solution Provider and ISV Accelerate Program, guidance on using AWS Co-sell Experience (ACE) and AWS Partner Funding Portal (APFP), and information on key topics, such as Generative AI and migration and modernization. These improvements provide Partners with structured, actionable information and quick access to resources, making it easier to navigate their AWS partnership. Partners can access guides in AWS Partner Central while completing assigned Tasks or by using Partner Assistant, the generative AI virtual assistant. From Task descriptions or Partner Assistant responses, partners can navigate directly to relevant guides for in-depth information and related resources. This integrated experience helps partners to find answers quickly, progress through their AWS Partner journey more efficiently and unlock benefits faster. Guides are available to all partners. To get started, log into AWS Partner Central, navigate to the «Resources» menu, and select «Guides». Learn more about becoming an AWS Partner.    

Today, Amazon Web Services (AWS) is announcing the general availability of notification actions in the AWS Console Mobile Application for iOS. Notification action buttons are displayed on the notification details screen when you receive a push notification from AWS User Notifications on your mobile device. AWS customers can take actions like view logs, start, stop, or reboot an EC2 instance on an event notification to shorten incident response diagnosis and resolution times.

With notification actions, you can quickly run common DevOps tasks based on your IAM permissions using notification action buttons. Notification actions are displayed automatically on notifications based on the state of the AWS resource. Using notification actions for event notifications, you can restart an EC2 instance the moment you receive an EC2 instance terminated push notification on your mobile device or get more information about an event notification without needing to return to your computer.

The Console Mobile App lets users view and manage a select set of resources to stay informed and connected with their AWS resources while on-the-go. Visit the product page for more information about the Console Mobile Application.

 

​Today, Amazon Web Services (AWS) is announcing the general availability of notification actions in the AWS Console Mobile Application for iOS. Notification action buttons are displayed on the notification details screen when you receive a push notification from AWS User Notifications on your mobile device. AWS customers can take actions like view logs, start, stop, or reboot an EC2 instance on an event notification to shorten incident response diagnosis and resolution times. With notification actions, you can quickly run common DevOps tasks based on your IAM permissions using notification action buttons. Notification actions are displayed automatically on notifications based on the state of the AWS resource. Using notification actions for event notifications, you can restart an EC2 instance the moment you receive an EC2 instance terminated push notification on your mobile device or get more information about an event notification without needing to return to your computer. The Console Mobile App lets users view and manage a select set of resources to stay informed and connected with their AWS resources while on-the-go. Visit the product page for more information about the Console Mobile Application.  

Today, AWS announces the release of Neuron 2.21, introducing support for AWS Trainium2 chips and Amazon EC2 Trn2 instances, including the trn2.48xlarge instance type and Trn2 UltraServer. This release also adds support for PyTorch 2.5 and introduces NxD Inference and Neuron Profiler 2.0 (beta). NxD Inference, is a new PyTorch-based library integrated with vLLM, simplifies the deployment of large language and multi-modality models and enables PyTorch model onboarding with minimal code changes, and Neuron Profiler 2.0 (beta), is new profiler that enhances capabilities and usability, including support for distributed workloads.

Neuron 2.21 also introduces Llama 3.1 405B model inference support using NxD Inference on a single trn2.48xlarge instance. The release updates Deep Learning Containers (DLCs) and Deep Learning AMIs (DLAMIs), and adds support for various model architectures, including Llama 3.2, Llama 3.3, and Mixture-of-Experts (MoE) models. New inference features include FP8 weight quantization and flash decoding for speculative decoding in Transformers NeuronX (TNx). Additionally, new training examples and features have been added, such as support for HuggingFace Llama 3/3.1 70B on Trn2 instances and DPO support for post-training model alignment.

AWS Neuron SDK supports training and deploying models on Trn1, Trn2, and Inf2 instances, available in AWS Regions as On-Demand Instances, Reserved Instances, Spot Instances, or part of Savings Plan.

For a full list of new features and enhancements in Neuron 2.21 and to get started with Neuron, see:

• AWS Neuron
• Trn2 Instances
• Trn1 Instances
• Inf2 Instances

 

​Today, AWS announces the release of Neuron 2.21, introducing support for AWS Trainium2 chips and Amazon EC2 Trn2 instances, including the trn2.48xlarge instance type and Trn2 UltraServer. This release also adds support for PyTorch 2.5 and introduces NxD Inference and Neuron Profiler 2.0 (beta). NxD Inference, is a new PyTorch-based library integrated with vLLM, simplifies the deployment of large language and multi-modality models and enables PyTorch model onboarding with minimal code changes, and Neuron Profiler 2.0 (beta), is new profiler that enhances capabilities and usability, including support for distributed workloads. Neuron 2.21 also introduces Llama 3.1 405B model inference support using NxD Inference on a single trn2.48xlarge instance. The release updates Deep Learning Containers (DLCs) and Deep Learning AMIs (DLAMIs), and adds support for various model architectures, including Llama 3.2, Llama 3.3, and Mixture-of-Experts (MoE) models. New inference features include FP8 weight quantization and flash decoding for speculative decoding in Transformers NeuronX (TNx). Additionally, new training examples and features have been added, such as support for HuggingFace Llama 3/3.1 70B on Trn2 instances and DPO support for post-training model alignment. AWS Neuron SDK supports training and deploying models on Trn1, Trn2, and Inf2 instances, available in AWS Regions as On-Demand Instances, Reserved Instances, Spot Instances, or part of Savings Plan. For a full list of new features and enhancements in Neuron 2.21 and to get started with Neuron, see:

AWS Neuron
Trn2 Instances
Trn1 Instances
Inf2 Instances