re:Invent 2024 Part 2 AI/ML 新功能發布

# re:Invent 2024 Part 2 AI/ML 新功能發布 Feature name, timestamp, and one paragraph introduction --- # Machine Learning and Generative AI ## AI/ML Managed Services :::info **Feature Name**: [AWS App Studio is now generally available](https://aws.amazon.com/about-aws/whats-new/2024/11/aws-app-studio-generally-available/) ✅ ::: Announced Nov 18, 2024 ⭐⭐⭐ Reference: - AWS Official Website: [Connect App Studio with Amazon S3](https://docs.aws.amazon.com/appstudio/latest/userguide/connectors-s3.html) :::success **Feature Name**: [Amazon Polly Long-Form Voices Expansion](https://aws.amazon.com/about-aws/whats-new/2024/11/three-long-form-voices/) ✔️ ::: Announced Nov 14, 2024 :::success **Feature Name**: [Amazon Polly launches more synthetic generative voices](https://aws.amazon.com/about-aws/whats-new/2024/11/amazon-polly-synthetic-generative-voices/) ✔️ ::: Announced Nov 21, 2024 Introduction: Amazon Polly has announced the general availability of seven highly expressive Generative voices in **English, French, Spanish, German, and Italian**, including Kajal, Bianca, Pedro, Andrés, Sergio, Daniel, and Rémi generative voices. :::success **Feature Name**: GenAI Index in Amazon Kendra ::: Announced Dec 4, 2024 Introduction: Amazon Kendra is an AI-powered search service enabling organizations to build intelligent search experiences and retrieval augmented generation (RAG) systems to power generative AI applications. Starting today, AWS customers can use a new index - the GenAI Index for RAG and intelligent search. With the Kendra GenAI Index, customers get high out-of-the-box search accuracy powered by the latest information retrieval technologies and semantic models. Reference: - AWS blog: [Introducing Amazon Kendra GenAI Index – Enhanced semantic search and retrieval capabilities](https://aws.amazon.com/blogs/machine-learning/introducing-amazon-kendra-genai-index-enhanced-semantic-search-and-retrieval-capabilities/) --- # Security **Feature Name**: [AWS PrivateLink - VPC Resources Access](https://aws.amazon.com/about-aws/whats-new/2024/12/access-vpc-resources-aws-privatelink/) Announced Dec 1, 2024 Introduction: AWS PrivateLink has announced a new capability that allows customers to use VPC endpoints to privately and securely access VPC resources. This enhancement enables access to resources such as databases or clusters that can be located in your VPC or on-premises network, without the requirement for load balancing. These resources can be shared with other teams within your organization or with external independent software vendor (ISV) partners. This marks an expansion from the previous functionality, which only permitted access to services using Network Load Balancer or Gateway Load Balancer, to now include any VPC resource that can be shared using AWS Resource Access Manager (AWS RAM). **Feature Name**: [AWS PrivateLink Cross-Region Connectivity](https://aws.amazon.com/about-aws/whats-new/2024/11/aws-privatelink-across-region-connectivity/) Announced November 26, 2024 Introduction: AWS PrivateLink now supports native cross-region connectivity. Until now, Interface VPC endpoints only supported connectivity to VPC endpoint services in the same region. This launch enables customers to connect to VPC endpoint services hosted in other AWS Regions in the same AWS partition over Interface endpoints. **Feature Name**: [AWS Declarative Policies](https://aws.amazon.com/about-aws/whats-new/2024/12/aws-declarative-policies/) Announced December 1, 2024 Introduction: AWS has announced the general availability of declarative policies, a new management policy type within AWS Organizations that simplifies how customers enforce durable intent for AWS service configurations across their organization. These policies prevent non-compliant actions and maintain configurations even when services add new APIs or features, or when customers add new principals or accounts to their organization, allowing customers to manage baseline configurations like EC2 instance launches and VPC public access controls through simple clicks or commands. **Feature Name**: [AWS Verified Access support for non-HTTP(S) protocols](https://aws.amazon.com/about-aws/whats-new/2024/12/aws-verified-access-secure-access-resources-non-https-protocols-preview/) Announced December 1, 2024 Introduction: AWS announces the preview of AWS Verified Access' new feature that supports secure access to resources that connect over protocols such as TCP, SSH, and, RDP. With this launch, Verified Access enables you to provide secure, VPN-less access to your corporate applications and resources using AWS zero trust principles. This feature eliminates the need to manage separate access and connectivity solutions for your non-HTTP(S) resources on AWS and simplifies security operations. --- # Resiliency **Feature Name**: [Amazon MemoryDB Multi-Region](https://aws.amazon.com/about-aws/whats-new/2024/12/general-availability-amazon-memory-db-multi-region/) Announced December 1, 2024 Introduction: AWS has announced the general availability of Amazon MemoryDB Multi-Region, a fully managed, active-active, multi-Region database service that enables building multi-Region applications with up to 99.999% availability and microsecond read and single-digit millisecond write latencies. This Valkey- and Redis OSS-compatible database service offers multi-AZ durability, asynchronous replication between regions within typically one second, and automatic resolution of update conflicts and data divergence issues, allowing customers to serve reads and writes locally from the Regions closest to their users while maintaining high performance and reliability. According to the announcement about Amazon MemoryDB Multi-Region, it is indeed designed for high availability, specifically: - It provides up to 99.999% availability - It offers active-active replication for increased resiliency - It allows serving reads and writes locally from the Regions closest to customers - It provides microsecond read and single-digit millisecond write latencies Regarding the technical difficulties in providing this technique, the document doesn't explicitly discuss the technical challenges. However, it mentions that MemoryDB Multi-Region handles some complex technical aspects automatically, such as: - Asynchronous replication between regions (typically within a second) - Automatic resolution of update conflicts - Automatic correction of data divergence issues --- # Simplicity **Feature Name**: [AWS Amplify AI kit for Amazon Bedrock](https://aws.amazon.com/about-aws/whats-new/2024/11/amazon-bedrock-prompt-management-available/) Announced November 19, 2024 Introduction: AWS has announced the general availability of the AWS Amplify AI kit for Amazon Bedrock, a comprehensive solution designed to help fullstack developers build web applications with AI capabilities including chat, conversational search, and summarization. This new kit enables developers with JavaScript or TypeScript knowledge to easily integrate AI experiences into their apps using Amazon Bedrock AI models, without requiring any prior machine learning expertise. The kit features a pre-built React UI component for streaming chat experiences, secure server-side access to Amazon Bedrock, and built-in capabilities for sharing user context with AI models. **Feature Name**: [High Performance Shared Cache for Mountpoint for Amazon S3](https://aws.amazon.com/about-aws/whats-new/2024/11/mountpoint-amazon-s3-high-performance-shared-cache/) Announced November 21, 2024 Introduction: Mountpoint for Amazon S3 now offers enhanced functionality by supporting Amazon S3 Express One Zone as a high performance read cache. This new capability allows the cache to be shared across multiple compute instances and can elastically scale to any dataset size. The feature enables Mountpoint for S3 to cache data in S3 Express One Zone after it's read, resulting in subsequent read requests being up to 7x faster compared to reading data from S3 Standard, making it particularly beneficial for compute-intensive use cases such as machine learning training for computer vision models where applications repeatedly read millions of small images from multiple instances. **Feature Name**: [Media Quality-Aware Resiliency (MQAR)](https://aws.amazon.com/about-aws/whats-new/2024/11/aws-announces-media-quality-aware-resiliency-live-streaming/) Announced November 21, 2024 Introduction: Media Quality-Aware Resiliency (MQAR) is a new integrated capability between Amazon CloudFront and AWS Media Services that provides dynamic, cross-region origin selection and failover based on a dynamically generated video quality score. This feature is designed for customers requiring always-on 'eyes-on-glass' to deliver live events and 24/7 programming channels, automatically switching between regions in seconds to recover from video quality degradation in one of the regions, ultimately helping to deliver a high quality of experience to viewers. **Feature Name**: [AWS Chatbot SDK](https://aws.amazon.com/about-aws/whats-new/2024/11/announcing-availability-aws-chatbot-sdk/) Announced November 22, 2024 Introduction: AWS has announced the general availability of AWS Chatbot SDKs, providing developers with access to AWS Chatbot's control plane APIs through the AWS SDK. This release enables customers to programmatically implement ChatOps workflows in their chat channels, allowing them to configure Microsoft Teams and Slack channels for monitoring and diagnosing issues, set up action buttons and command aliases for quick telemetry retrieval, and programmatically tag resources for enforcing tag-based controls in their environments. **Feature Name**: [Storage Browser for Amazon S3](https://aws.amazon.com/about-aws/whats-new/2024/12/storage-browser-amazon-s3/) Announced December 1, 2024 Introduction: AWS has announced the general availability of Storage Browser for S3, an open source component that can be integrated into web applications to provide end users with a simple interface for managing data stored in S3. This component, available in AWS Amplify React and JavaScript client libraries, allows authorized users to browse, download, upload, search, copy, and delete data in S3 directly from applications, while incorporating new security features such as automatic checksum calculations to ensure data durability. Reference: - [Announcing AWS Transfer Family web apps](https://aws.amazon.com/about-aws/whats-new/2024/12/aws-transfer-family-web-apps/) **Feature Name**: [Amazon S3 Default Data Integrity Protections](https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-s3-default-data-integrity-protections/) Announced December 1, 2024 Introduction: Amazon S3 has introduced new default data integrity protections for object upload requests, where the latest AWS SDKs automatically calculate CRC-based checksums for uploads during network transmission. S3 verifies these checksums and only accepts objects after confirming data integrity was maintained during transit over the public internet. Additionally, S3 now stores a CRC-based whole-object checksum in object metadata, even for multipart uploads, enabling users to verify object integrity stored in S3 at any time. **Feature Name**: [Amazon CloudWatch and Amazon OpenSearch Service Integration](https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-cloudwatch-opensearch-service-integrated-analytics/) Announced December 1, 2024 Introduction: Amazon Web Services has launched a new integrated analytics experience and zero-ETL integration between Amazon CloudWatch and Amazon OpenSearch Service, allowing customers to leverage both services' capabilities. CloudWatch customers can now use OpenSearch's Piped Processing Language (PPL) and OpenSearch SQL, while also benefiting from out-of-the-box curated dashboards for vended logs. OpenSearch customers can analyze CloudWatch Logs without data duplication. **Feature Name**: [AWS DMS Schema Conversion with generative AI](https://aws.amazon.com/about-aws/whats-new/2024/12/aws-dms-schema-conversion-generative-ai/) Announced Dec 1, 2024 Introduction: AWS Database Migration Service (AWS DMS) Schema Conversion with generative AI is now available. The feature is currently available for database schema conversion from commercial engines, such as Microsoft SQL Server, to Amazon Aurora PostgreSQL-Compatible Edition and Amazon Relational Database Service (Amazon RDS) for PostgreSQL. **Feature Name**: [AWS Step Functions Variables and JSONata transformations](https://aws.amazon.com/about-aws/whats-new/2024/11/aws-step-functions-variables-jsonata-transformations/) Announced Nov 22, 2024 Introduction: AWS Step Functions has introduced two new capabilities: Variables and JSONata data transformations. Variables enable developers to assign data in one state and reference it in a subsequent state, simplifying state payload management and reducing the need to pass data through multiple intermediate states. JSONata support, an open source query and transformation language, allows customers to perform advanced data manipulation and transformation including date and time formatting, and mathematical operations. These features are designed to simplify input and output processing by reducing the number of JSON transformation fields required to call services and pass data to the next state. **Feature Name**: [AWS Chatbot Integration with Amazon Bedrock Agents](https://aws.amazon.com/about-aws/whats-new/2024/09/aws-chatbot-amazon-bedrock-agent-microsoft-teams-slack/) Announced September 17, 2024 Introduction: AWS Chatbot has introduced a new capability that enables customers to interact with Amazon Bedrock agents directly from Slack and Microsoft Teams chat channels. This integration eliminates the need for developing custom chat applications, as users can now simply connect their Bedrock agent alias with an AWS Chatbot channel configuration and interact with the agent by tagging it in their chat channels using commands like "@aws ask connectorname ...". The feature is available at no additional cost in all AWS Regions where AWS Chatbot is offered. --- # Performance **Feature Name**: Disk-optimized Vector Engine for Amazon OpenSearch Service Announced November 19, 2024 Introduction: Amazon OpenSearch Service has introduced a disk-optimized vector engine that enables running modern search applications at one-third of the cost on OpenSearch 2.17 domains. This new capability allows k-NN (vector) indexes to be optimized for low memory environments through disk mode configuration, utilizing compression techniques like binary quantization while maintaining search quality through a disk-optimized rescoring mechanism with full-precision vectors. This cost-efficient solution is ideal for vector search workloads that require high accuracy and can accommodate latencies in the low hundreds of milliseconds range, offering a more economical alternative to the existing in-memory mode when single-digit latency isn't necessary. **Feature Name**: [Amazon MSK Vector Embedding Generation with Amazon Bedrock](https://aws.amazon.com/about-aws/whats-new/2024/11/amazon-msk-vector-embedding-generation-amazon-bedrock/) Announced November 6, 2024 Introduction: Amazon MSK (Managed Streaming for Apache Kafka) now supports new Managed Streaming for Apache Flink blueprints to generate vector-embeddings using Amazon Bedrock, enabling easier development of real-time AI applications powered by up-to-date, contextual data. This new feature simplifies the process of incorporating the latest data from Amazon MSK streaming pipelines into generative AI models, eliminating the need for custom code to integrate real-time data streams, vector databases, and large language models. --- # Cost and Monitoring **Feature Name**: [Cost analysis capabilities in Amazon Q Developer](https://aws.amazon.com/about-aws/whats-new/2024/11/amazon-q-developer-natural-language-cost-analysis/) Announced November 26, 2024 ⭐⭐⭐ Introduction: AWS has announced the addition of cost analysis capabilities to Amazon Q Developer, enabling customers to retrieve and interpret their AWS cost data through natural language interactions. This new feature allows users of all skill levels to better understand and manage their AWS spending without requiring previous knowledge of AWS Cost Explorer by asking natural language questions about their AWS costs, with Q interpreting these questions, analyzing the relevant cost data, and providing easy-to-understand responses. **Feature Name**: [Amazon CloudWatch Database Insights](https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-cloudwatch-database-insights/) Announced December 1, 2024 Introduction: Amazon CloudWatch Database Insights is a database observability solution that provides a curated experience for DevOps engineers, application developers, and DBAs to expedite database troubleshooting and gain a holistic view of database fleet health. The solution consolidates logs and metrics from applications, databases, and operating systems into a unified console view, offering pre-built dashboards, recommended alarms, and automated telemetry collection. It supports Amazon Aurora PostgreSQL and Amazon Aurora MySQL, allowing users to monitor fleet health and perform guided troubleshooting for root-cause analysis, while enabling application developers to correlate database dependencies with application performance through integration with CloudWatch Application Signals. **Feature Name**: [AWS Invoice Configuration](https://aws.amazon.com/about-aws/whats-new/2024/12/aws-invoice-configuration/) Announced December 1, 2024 Introduction: AWS has announced the general availability of Invoice Configuration, a new feature that enables customers to customize their invoicing experience by receiving separate AWS invoices based on their organizational structure. This feature allows customers to group AWS accounts according to their internal business entities such as legal entities, subsidiaries, and cost centers, enabling them to receive separate AWS invoices for each business entity within the same AWS Organization, thus streamlining invoice processing by eliminating manual processes for invoice splitting at the entity level. **Feature Name**: [Enhanced AWS Pricing Calculator with Discounts and Purchase Commitments](https://aws.amazon.com/about-aws/whats-new/2024/11/aws-pricing-calculator-estimates-purchase-commitments-discounts/) Announced November 25, 2024 Introduction: AWS has announced the public preview of an enhanced AWS Pricing Calculator that incorporates eligible discounts and helps estimate the cost impact of commitment purchases on consolidated bills. The enhanced calculator, available within the AWS Billing and Cost Management Console, provides two types of cost estimates: workload cost estimation and full AWS bill estimation. Users can import historical usage, create new usage estimates, import existing Pricing Calculator estimates, and share estimates with other AWS console users, enabling informed financial planning decisions for workload migrations and growth planning. **Feature Name**: [Amazon Q account resources chat in AWS Console Mobile App](https://aws.amazon.com/about-aws/whats-new/2024/11/aws-announces-amazon-q-account-console-mobile-app/) Announced November 22, 2024 Introduction: Amazon Web Services (AWS) has announced the general availability of Amazon Q Developer's AWS account resources chat capability in the AWS Console Mobile Application. This new feature allows users to utilize their device's voice input and output capabilities along with natural language prompts to interact with their AWS resources while on-the-go, enabling them to list resources in their AWS account, get specific resource details, and inquire about related resources through simple voice or text commands. **Feature Name**: [Amazon CloudWatch Logs field indexes and enhanced log group selection in Logs Insights](https://aws.amazon.com/about-aws/whats-new/2024/11/amazon-cloudwatch-logs-field-indexes-log-group-selection-log-insights/) Announced Nov 21, 2024 Introduction: Amazon CloudWatch Logs introduces field indexes and enhanced log group selection to accelerate log analysis. Now, you can index critical log attributes like requestId and transactionId to accelerate query performance and scan relevant indexed data. This means faster troubleshooting, and easier identification of trends. You can create up to 20 field indexes per log group, and once defined, all future logs matching the defined fields will remain indexed for up to 30 days. Additionally, CloudWatch Logs Insights now supports querying up to 10,000 log groups, across one or more accounts linked via cross-account observability. **Feature Name**: [Amazon CloudWatch Application Signals enhanced search and analytics for application transactions](https://aws.amazon.com/about-aws/whats-new/2024/11/amazon-cloudwatch-visibility-application-transactions/) Announced Nov 21, 2024 Introduction: AWS announces the general availability of an enhanced search and analytics experience in CloudWatch Application Signals. This feature empowers developers and on-call engineers with complete visibility into application transaction spans, which are the building blocks of distributed traces that capture detailed interactions between users and various application components. **Feature Name**: [Amazon Q Business Analytics Dashboard and CloudWatch Logs Integration]() Announced Oct 22, 2024 Introduction: Amazon Q Business has introduced a new analytics dashboard and Amazon CloudWatch Logs integration, providing administrators with comprehensive insights into their Amazon Q Business application environments and Amazon Q Apps. The analytics dashboard offers interactive charts and visualizations for monitoring key metrics, while the CloudWatch Logs integration enables detailed analysis of conversation data and user feedback, with data also available in Amazon S3 and Amazon Data Firehouse. # Developer **Feature Name**: [Amazon Q Developer transformation capabilities for mainframe modernization](https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-q-developer-transformation-mainframe-modernization/) Announced December 3, 2024 Introduction: AWS announces new generative AI–powered capabilities of Amazon Q Developer in public preview to help customers and partners accelerate large-scale assessment and modernization of mainframe applications. Amazon Q Developer agents can autonomously refactor COBOL code into cloud-optimized Java code while preserving business logic. **Feature Name**: [Amazon Q Developer Code Review Automation](https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-q-developer-automate-code-reviews/) Announced December 3, 2024 Amazon Q Developer has introduced automated code review capabilities, allowing developers to receive automatic comments on their code within their IDE, identify suspicious code patterns, access available patches, and assess deployment risks for faster feedback. This new feature is accessible through a /review chat command in supported IDEs (Visual Studio Code and IntelliJ IDEA) and is available with both Free Tier and Pro Tier subscriptions of Amazon Q Developer. **Feature Name**: [Amazon Q Developer Documentation Generation](https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-q-developer-generate-documentation-source-code/) Announced December 3, 2024 Introduction: Starting today, Amazon Q Developer can document your code by automatically generating readme files and data-flow diagrams within your projects. This new capability addresses the fact that developers spend an average of just one hour per day coding, with most of their time consumed by tedious tasks like documentation and codebase learning. Q Developer, as a generative AI-powered assistant, can now help developers understand existing code bases faster and quickly document new features, allowing them to focus more on shipping features for customers.