Top takeaways from AWS re:Invent 2024

From AI innovations to streamlining security and managing cloud complexity, AWS re:Invent 2024 was filled with product announcements and thought-provoking insights. 

If you attended the conference and need a refresher or couldn’t make it and want the quick highlights, we’ve rounded up the top takeaways from the event.

Want a full recap? Check out our 2024 AWS re:Invent coverage:

• AWS re:Invent 2024: SVP Peter DeSantis' keynote highlights
• AWS re:Invent 2024: CEO Matt Garman’s keynote highlights
• AWS re:Invent 2024: VP Dr. Swami Sivasubramanian’s keynote
• AWS re:Invent 2024: Dr. Werner Vogels’ Keynote Highlights
• The AWS re:Invent 2024 swag review: The best of the best

By far one of the most exciting announcements this year was the preview of Amazon Aurora DSQL. In AWS’s words, this is a new serverless, distributed SQL database with active-active high availability. As a result, you can build applications with immense scalability, the highest availability, and zero infrastructure management.

The new SageMaker Lakehouse unifies data across your Amazon S3 data lakes and Amazon Redshift data warehouses. By removing data siloes and reducing duplicate data, Lakehouse makes it easier to build comprehensive analytics and AI/ML applications.

As organizations look to scale their AI and ML models, they need the compute power to support them. Amazon’s new EC2 Trn2 instances are 4x faster, with 4x more memory bandwidth and 3x more memory capacity than the first-generation Trn1 instances.

The new UltraServers provide 64x Trainium2 chips with high-bandwidth, low-latency NeuronLink interconnect. Together, these new services offer scaled up compute power for ML and generative AI training performance.

AWS unveiled Amazon Q Developer at re:Invent 2023. This year, they’ve expanded the AI-powered agent’s capabilities to further improve developer productivity. 

Amazon Q Developer can now generate documentation, such as data flow diagrams, in codebases. You can also use it to automate code reviews and find and fix code quality issues faster. Lastly, it can automate the unit test process to identify test cases, write unit tests, and generate basic cases.

Amazon SageMaker is now Amazon SageMaker AI. The renamed platform is meant to be a single hub for:

• Data exploration, preparation and integration
• Big data processing
• SQL analytics
• ML model development and training
• Generative AI application development

Amazon Nova is a new collection of AI foundational models designed to help developers generate creative content, process images and text, build AI assistants, and more, all while reducing costs and latency. There will be six total Nova models, each with different focuses to meet different needs.

Anyone who’s gone through the procurement process knows how lengthy it can be. AWS aims to streamline that process with their new Buy with AWS program. This allows you to discover and purchase solutions available in the AWS Marketplace from AWS Partner sites.

Find Pluralsight on the AWS Marketplace.

If you run container workloads on Elastic Container Service (ECS), the new Container Insights will give you enhanced observability, making it easier to prevent, identify, and resolve issues.

Container Insights can also help you:

• Manage ECS resources with dashboards
• Track deployments and root causes of deployment failures
• Monitor resources across accounts

AWS Database Migration Service Schema Conversion (AWS DMS SC) now uses generative AI to automate some of the most time-intensive schema conversion tasks. By automating more of the schema conversion process when migrating between databases, you can reduce migration costs and speed up project timelines.

The new AWS Security Incident Response service aims to help users respond to security events faster and more efficiently. It does this through three key features: automating triage and investigation, offering preconfigured notification rules and permission settings for incident response, and providing self-service investigation tools and 24/7 support from the AWS Customer Incident Response Team (CIRT).

For storing Tabular data at scale, the new Amazon S3 Tables provide cloud object storage with built-in Apache Iceberg support. These tables are designed for analytics workloads, with up to 3x faster query throughput and up to 10x higher transactions per second compared to self-managed tables.

Likewise, table buckets are meant to store tabular data, empowering you to quickly create tables and manage access with table-level permissions.

Is it a challenge to find objects in your S3 buckets? When S3 objects are added or modified, metadata will automatically be generated and stored in Apache Iceberg tables. While still in preview, this feature will make it much easier to find the data you need for analytics, data processing, AI training, and other needs.

Growth and innovation often go hand in hand with complexity. Dr. Werner Vogels addressed that in his keynote. But he also gave advice on how to manage it by preventing unintended complexity and planning for intended complexity. 

To make operations as simple as possible, he recommends automating as much as possible, working with smaller teams to limit risk, and building infrastructure that can evolve, rather than constantly adding a patchwork of new features.

Amid the new product announcements and emphasis on innovation, AWS continues to call for a focus on the fundamentals. Only with a strong foundation will you be able to make the most of new technologies.

Start building your AWS skills:

• AWS Solutions Architect Learning Path
• Serverless Development on AWS
• Hands-On Labs: Building and Deploying with AWS Lambda
• Hands-On Labs: Train a ML model with SageMaker