What’s New From MongoDB at AWS re:Invent 2024

As thousands of attendees make their way home after a week in Vegas—a week packed with learning, product launches, and round-the-clock events—we thought we’d reflect on the show’s highlights. MongoDB was excited to showcase our latest integrations and solutions with Amazon Web Services (AWS), which range from new ways to optimize generative AI, to faster, more cost-effective methods for modernizing applications.

But first, we want to thank our friends at AWS for recognizing MongoDB as the AWS Technology Partner of the Year NAMER! This prestigious award recognizes AWS Technology Partners that are using AWS to lower costs, increase agility, and innovate faster. Announced during the re:Invent Partner Awards Gala, the Technology Partner of the Year Award is a testament to the specialization, innovation, and cooperation MongoDB and AWS have jointly brought to customers this year.

In addition, MongoDB also received AWS Partner of the Year awards for Italy, Turkey, and Iberia. These awards follow wins in the ASEAN Global Software Partner of the Year and Taiwan Technology Partner of the Year categories earlier in the year, further demonstrating the global reach and popularity of the world’s most popular document database!

Harnessing the potential of gen AI

If 2024 (and 2023, and 2022…) was the year of gen AI excitement, then 2025 may turn out to be marked by realistic gen AI implementation. Indeed, we’ve seen customers shift their 2025 AI focus toward optimizing resource-intensive gen AI workloads to drive down costs—and to get the most out of this groundbreaking technology.

Retrieval-augmented generation (RAG), one of the main ways companies use their data to customize the output of foundation models, has become the focus of this push for optimization. Customers are looking for easier ways to fine-tune their RAG systems, asking questions like, “How do I evaluate the efficiency and accuracy of my current RAG workflow?”

To that end, AWS and MongoDB are introducing new services and technologies for enterprises to optimize RAG architecture compute costs, while also maintaining accuracy.

First up is vector quantization. By reducing vector storage and memory requirements while preserving performance, these capabilities empower developers to build AI-enriched applications with more scale—and at a lower cost. Leading foundation models like Amazon Titan are already compatible with vector quantization, helping to maintain high accuracy of generated responses while simultaneously reducing costs. You can read more about vector quantization on our blog.

As for RAG evaluation, AWS has launched a new feature for Amazon Bedrock called, naturally, RAG Evaluator. This tool allows Bedrock users to evaluate and monitor RAG Apps natively within the Bedrock environment, eliminating the need for third-party frameworks to run tests and comparisons.

As a Knowledge Base for Amazon Bedrock, MongoDB Atlas is ready on day one to take advantage of Bedrock RAG Evaluator, allowing companies to gauge and compare the quality of their RAG Apps across different applications.

The RAG Evaluator, built on several joint integrations and solutions AWS and MongoDB released in 2024, and vector quantization together can streamline the deployment of enterprise generative AI.

For example, in October MongoDB, Anthropic, and AWS announced a joint solution to create a memory-enhanced AI agent. Together, the three partners offer enterprise-grade, trusted, secure technologies to build generative AI apps quickly and flexibly using a family of foundation models in a fully managed environment.

Overall, MongoDB and AWS are making it easier—and more cost-effective—for developers to build innovative applications that harness the full potential of generative AI on AWS.

From cars to startups to glue

MongoDB and AWS have been hard at work on a number of other solutions for developers across industries. Here’s a quick roundup:

AWS Amplify + AppSync + MongoDB

For startups, or for any organization looking to quickly test and launch applications, speed is everything. That’s why MongoDB teamed up with AWS to create a full-stack solution that provides developers with the same high standards of performance and scalability they would demand for any app.

By combining AWS Amplify, AWS AppSync, and MongoDB Atlas, AWS and MongoDB have created a full-stack solution that enables seamless front-end development, robust and scalable backend services, out-of-the-box CI/CD, and a flexible and powerful database solution, allowing developers to drastically reduce the coding time required to launch new applications. Check out this tutorial and repository for a starter template.

Digital twins on AWS CMS

For those in the automotive sector, MongoDB and AWS have developed a connected mobility solution to help remove the undifferentiated integration, or “technical plumbing” work, of connecting vehicles to the cloud.

When used together, Connected Mobility Solution (CMS) on AWS and MongoDB Atlas help accelerate the development of next-generation digital twin use cases and applications, including connected car use cases.

MongoDB’s document model allows easy and flexible modeling and storage of connected vehicle sensor data. Read our joint blog with AWS to learn how the MongoDB Atlas document model helps with data modeling of connected vehicles data and how this capability can be leveraged via AWS Automotive Cloud Developer Portal (ACDP).

AWS Glue + MongoDB Atlas

Speaking of undifferentiated plumbing, MongoDB Atlas is now integrated into AWS Glue’s visual interface. The new integration simplifies data integration between MongoDB Atlas and AWS, making it easy to build efficient ETL (Extract, Transform, Load) pipelines with minimal effort.

With its visual interface, AWS Glue allows users to seamlessly transfer, transform, and load data to and from MongoDB Atlas without needing deep technical expertise in Spark or SQL.

In this blog post, we look at how AWS Glue and MongoDB Atlas can transform the way you manage data movement.

Buy with AWS

In the spirit of making things easier for our joint customers, in early 2025 MongoDB will also join the AWS ‘Buy with AWS’ program.

Once up and running, Buy With AWS will allow customers to pay for Atlas using their AWS account directly from the Atlas UI, further reducing friction for customers wanting to get started with Atlas on AWS.

New Atlas Updates Announced at re:Invent

Aside from our joint endeavors with AWS, MongoDB has also been hard at work on improving the core Atlas platform. Here’s an overview of what we announced:

Asymmetrical sharding support for Terraform Atlas Provider

Customers are constantly seeking ways to optimize costs to ensure they get the best value for their resources. With Asymmetrical Sharding, now available in the Terraform MongoDB Atlas Provider, MongoDB Atlas users can customize the Cluster Tier and IOPS for each shard, encouraging better resource allocation, improved operational efficiency, and cost savings as customer needs evolve.

Atlas Flex Tier

Our new Atlas Flex tier offers the scaled flexibility of serverless, with the cost-capped assurance of shared tier clusters. With Atlas Flex Tier, developers can build and scale applications cost-effectively without worrying about runaway bills or resource provisioning.

New test bench feature in Query Converter

At MongoDB, we firmly believe that the document model is the best way for customers to build applications with their data. In our latest update to Relational Migrator, we’ve introduced Generative AI to automatically convert SQL database objects and validate them using the test bench in a fraction of the time, producing deployment-ready code up to 90% faster.

This streamlined approach reduces migration risks and manual development effort, enabling fast, efficient, and precise migrations to MongoDB.

Source: Read More