12/6/2024, 12:00:00 AM ~ 12/9/2024, 12:00:00 AM (UTC)

Recent Announcements

Amazon EC2 Hpc6id instances are now available in Europe (Paris) region

Starting today, Amazon EC2 Hpc6id instances are available in additional AWS Region Europe (Paris). These instances are optimized to efficiently run memory bandwidth-bound, data-intensive high performance computing (HPC) workloads, such as finite element analysis and seismic reservoir simulations. With EC2 Hpc6id instances, you can lower the cost of your HPC workloads while taking advantage of the elasticity and scalability of AWS.\n EC2 Hpc6id instances are powered by 64 cores of 3rd Generation Intel Xeon Scalable processors with an all-core turbo frequency of 3.5 GHz, 1,024 GB of memory, and up to 15.2 TB of local NVMe solid state drive (SSD) storage. EC2 Hpc6id instances, built on the AWS Nitro System, offer 200 Gbps Elastic Fabric Adapter (EFA) networking for high-throughput inter-node communications that enable your HPC workloads to run at scale. The AWS Nitro System is a rich collection of building blocks that offloads many of the traditional virtualization functions to dedicated hardware and software. It delivers high performance, high availability, and high security while reducing virtualization overhead. To learn more about EC2 Hpc6id instances, see the product detail page.

Amazon EC2 Hpc7a instances are now available in Europe (Paris) region

Starting today, Amazon EC2 Hpc7a instances are available in additional AWS Region Europe (Paris). EC2 Hpc7a instances are powered by 4th generation AMD EPYC processors with up to 192 cores, and 300 Gbps of Elastic Fabric Adapter (EFA) network bandwidth for fast and low-latency internode communications. Hpc7a instances feature Double Data Rate 5 (DDR5) memory, which enables high-speed access to data in memory.\n Hpc7a instances are ideal for compute-intensive, tightly coupled, latency-sensitive high performance computing (HPC) workloads, such as computational fluid dynamics (CFD), weather forecasting, and multiphysics simulations, helping you scale more efficiently on fewer nodes. To optimize HPC instances networking for tightly coupled workloads, you can access these instances in a single Availability Zone within a Region. To learn more, see Amazon Hpc7a instances.

Amazon Aurora now available as a quick create vector store in Amazon Bedrock Knowledge Bases

Amazon Aurora PostgreSQL is now available as a quick create vector store in Amazon Bedrock Knowledge Bases. With the new Aurora quick create option, developers and data scientists building generative AI applications can select Aurora PostgreSQL as their vector store with one click to deploy an Aurora Serverless cluster preconfigured with pgvector in minutes. Aurora Serverless is an on-demand, autoscaling configuration where capacity is adjusted automatically based on application demand, making it ideal as a developer vector store.\n Knowledge Bases securely connects foundation models (FMs) running in Bedrock to your company data sources for Retrieval Augmented Generation (RAG) to deliver more relevant, context-specific, and accurate responses that make your FM more knowledgeable about your business. To implement RAG, organizations must convert data into embeddings (vectors) and store these embeddings in a vector store for similarity search in generative artificial intelligence (AI) applications. Aurora PostgreSQL, with the pgvector extension, has been supported as a vector store in Knowledge Bases for existing Aurora databases. With the new quick create integration with Knowledge Bases, Aurora is now easier to set up as a vector store for use with Bedrock. The quick create option in Bedrock Knowledge Bases is available in these regions with the exception of AWS GovCloud (US-West) which is planned for Q4 2024. To learn more about RAG with Amazon Bedrock and Aurora, see Amazon Bedrock Knowledge Bases. Amazon Aurora combines the performance and availability of high-end commercial databases with the simplicity and cost-effectiveness of open-source databases. To get started using Amazon Aurora PostgreSQL as a vector store for Amazon Bedrock Knowledge Bases, take a look at our documentation.

Amazon CloudWatch now provides centralized visibility into telemetry configurations

Amazon CloudWatch now offers centralized visibility into critical AWS service telemetry configurations, such as Amazon VPC Flow Logs, Amazon EC2 Detailed Metrics, and AWS Lambda Traces. This enhanced visibility enables central DevOps teams, system administrators, and service teams to identify potential gaps in their infrastructure monitoring setup. The telemetry configuration auditing experience seamlessly integrates with AWS Config to discover AWS resources, and can be turned on for the entire organization using the new AWS Organizations integration with Amazon CloudWatch.\n With visibility into telemetry configurations, you can identify monitoring gaps that might have been missed in your current setup. For example, this helps you identify gaps in your EC2 detailed metrics so that you can address them and easily detect short-lived performance spikes and build responsive auto-scaling policies. You can audit telemetry configuration coverage at both resource type and individual resource levels, refining the view by filtering across specific accounts, resource types, or resource tags to focus on critical resources. The telemetry configurations auditing experience is available in US East (N. Virginia), US West (Oregon), US East (Ohio), Asia Pacific (Tokyo), Asia Pacific (Singapore), Asia Pacific (Sydney), Europe (Frankfurt), Europe (Ireland), and Europe (Stockholm) regions. There is no additional cost to turn on the new experience, including for AWS Config. You can get started with auditing your telemetry configurations using the Amazon CloudWatch Console, by clicking on Telemetry config in the navigation panel, or programmatically using the API/CLI. To learn more, visit our documentation.

AWS Config now supports a service-linked recorder

AWS Config added support for a service-linked recorder, a new type of AWS Config recorder that is managed by an AWS service and can record configuration data on service-specific resources, such as the new Amazon CloudWatch telemetry configurations audit. By enabling the service-linked recorder in Amazon CloudWatch, you gain centralized visibility into critical AWS service telemetry configurations, such as Amazon VPC Flow Logs, Amazon EC2 Detailed Metrics, and AWS Lambda Traces.\n With service-linked recorders, an AWS service can deploy and manage an AWS Config recorder on your behalf to discover resources and utilize the configuration data to provide differentiated features. For example, an Amazon CloudWatch managed service-linked recorder helps you identify monitoring gaps within specific critical resources within your organization, providing a centralized, single-pane view of telemetry configuration status. Service-linked recorders are immutable to ensure consistency, prevention of configuration drift, and simplified experience. Service-linked recorders operate independently of any existing AWS Config recorder, if one is enabled. This allows you to independently manage your AWS Config recorder for your specific use cases while authorized AWS services can manage the service-linked recorder for feature specific requirements. Amazon CloudWatch managed service-linked recorder is now available in US East (N. Virginia), US West (Oregon), US East (Ohio), Asia Pacific (Tokyo), Asia Pacific (Singapore), Asia Pacific (Sydney) Europe (Frankfurt), Europe (Ireland), Europe (Stockholm) regions. The AWS Config service-linked recorder specific to Amazon CloudWatch telemetry configuration feature is available to customers at no additional cost. To learn more, please refer to our documentation.

Amazon RDS Performance Insights extends On-demand Analysis to new regions

Amazon RDS (Relational Database Service) Performance Insights expands the availability of its on-demand analysis experience to 15 new regions. This feature is available for Aurora MySQL, Aurora PostgreSQL, and RDS for PostgreSQL engines.\n This on-demand analysis experience, which was previously available in only 15 regions, is now available in all commercial regions. This feature allows you to analyze Performance Insights data for a time period of your choice. You can learn how the selected time period differs from normal, what went wrong, and get advice on corrective actions. Through simple-to-understand graphs and explanations, you can identify the chief contributors to performance issues. You will also get the guidance on the next steps to act on these issues. This can reduce the mean-time-to-diagnosis for database performance issues from hours to minutes. Amazon RDS Performance Insights is a database performance tuning and monitoring feature of RDS that allows you to visually assess the load on your database and determine when and where to take action. With one click in the Amazon RDS Management Console, you can add a fully-managed performance monitoring solution to your Amazon RDS database. To learn more about RDS Performance Insights, read the Amazon RDS User Guide and visit Performance Insights pricing for pricing details and region availability.

SageMaker SDK enhances training and inference workflows

Today, we are introducing the new ModelTrainer class and enhancing the ModelBuilder class in the SageMaker Python SDK. These updates streamline training workflows and simplify inference deployments.\n The ModelTrainer class enables customers to easily set up and customize distributed training strategies on Amazon SageMaker. This new feature accelerates model training times, optimizes resource utilization, and reduces costs through efficient parallel processing. Customers can smoothly transition their custom entry points and containers from a local environment to SageMaker, eliminating the need to manage infrastructure. ModelTrainer simplifies configuration by reducing parameters to just a few core variables and providing user-friendly classes for intuitive SageMaker service interactions. Additionally, with the enhanced ModelBuilder class, customers can now easily deploy HuggingFace models, switch between developing in local environment to SageMaker, and customize their inference using their pre- and post-processing scripts. Importantly, customers can now pass the trained model artifacts from ModelTrainer class easily to ModelBuilder class, enabling a seamlessly transition from training to inference on SageMaker. You can learn more about ModelTrainer class here, ModelBuilder enhancements here, and get started using ModelTrainer and ModelBuilder sample notebooks.

Amazon SageMaker introduces new capabilities to accelerate scaling of Generative AI Inference

We are excited to announce two new capabilities in SageMaker Inference that significantly enhance the deployment and scaling of generative AI models: Container Caching and Fast Model Loader. These innovations address critical challenges in scaling large language models (LLMs) efficiently, enabling faster response times to traffic spikes and more cost-effective scaling. By reducing model loading times and accelerating autoscaling, these features allow customers to improve the responsiveness of their generative AI applications as demand fluctuates, particularly benefiting services with dynamic traffic patterns.\n Container Caching dramatically reduces the time required to scale generative AI models for inference by pre-caching container images. This eliminates the need to download them when scaling up, resulting in significant reduction in scaling time for generative AI model endpoints. Fast Model Loader streams model weights directly from Amazon S3 to the accelerator, loading models much faster compared to traditional methods. These capabilities allow customers to create more responsive auto-scaling policies, enabling SageMaker to add new instances or model copies quickly when defined thresholds are reached, thus maintaining optimal performance during traffic spikes while at the same time managing costs effectively. These new capabilities are accessible in all AWS regions where Amazon SageMaker Inference is available. To learn more see our documentation for detailed implementation guidance.

AWS Blogs

AWS Japan Blog (Japanese)

AWS Architecture Blog

AWS Database Blog

AWS for Industries

AWS Machine Learning Blog

AWS Security Blog

Open Source Project

AWS CDK