MongoDB Atlas AWS
Foundational models (FMs) are trained on vast volumes of data using billions of parameters. However, in order to reply to client requests about domain-specific private data, they have to consult a trustworthy knowledge base that is separate from the model's training data sources. One method that is often used to do this is Retrieval Augmented Generation (RAG). By obtaining data from internal or business sources, RAG broadens the scope of FMs' capabilities without necessitating a retraining of the model. It is a reasonably priced way to improve model output such that it is relevant, accurate, and useful in a range of contexts.
Knowledge Bases for Amazon Bedrock is a fully managed functionality that helps with the deployment of the entire RAG workflow, from intake to retrieval and fast augmentation, without requiring the creation of unique connections for data sources and managing data flows.
AWS has announced that MongoDB Atlas is now available as a vector store in Knowledge Bases for Amazon Bedrock. Using a MongoDB Atlas vector store connection, you may develop RAG solutions to securely connect FMs in Amazon Bedrock to your business's private data sources. Knowledge Bases for Amazon Bedrock now support the vector engines for Pinecone, Redis Enterprise Cloud, Amazon Aurora PostgreSQL-Compatible Edition, and Amazon OpenSearch Serverless.
RAG apps are created with MongoDB Atlas and Amazon Bedrock Knowledge Bases
Vector search in MongoDB Atlas is powered by the vectorSearch index type. The index specification needs to include the vector type of the field that holds the vector data. To use MongoDB Atlas vector search in your application, you must first create an index, ingest source data, create vector embeddings, and store them in a MongoDB Atlas collection. You have to first transform the input text into a vector embedding before you can start running searches. Following that, you may use an aggregate pipeline stage to run vector search queries against fields that are indexed as the vector type in a vectorSearch type index.
The MongoDB Atlas connection with Knowledge Bases for Amazon Bedrock manages most of the labor-intensive operations. Once the knowledge base and vector search index are configured, you may include RAG into your applications. Amazon Bedrock will convert your input (prompt) into embeddings, query the knowledge base, add contextual data from the search results to the FM prompt, and return the resultant response.
What is Atlas MongoDB?
The most advanced cloud database solution available, with unmatched data distribution and mobility across AWS, Azure, and Google Cloud, as well as integrated automation for workload and resource optimisation.
MongoDB Atlas is an integrated suite of data services that simplifies and accelerates the process of developing with data, centred around a cloud database. You may create more rapidly and wisely if you have a developer data platform that helps you overcome your data challenges.
Utilise applications everywhere
To run anywhere in the world, use Atlas. You may launch a database in more than 90 regions using AWS, Azure, and Google Cloud, and expand it to be global, multi-regional, or multi-cloud as needed. Pin data to specified places for extremely reduced latency and strict compliance requirements.
Scale up operations with confidence
Build with confidence. Atlas is pre-configured with best practices and intelligently automates the operations required to ensure the security of your data and the proper operation of your database.
Reduce the architecture's complexity
Access and query your data for any use case with only one query API. The entire AWS platform can instantly access data stored in Atlas, including full-text search, analytics, and visualisations.
Take note of the delivery specifications
Keep your apps running even if traffic triples or new features are added. Atlas comes with advanced speed optimisation features to make sure you always have the database resources you need to keep producing.
MongoDB Atlas on AWS
Create intelligent, enterprise-ready apps by combining AWS with MongoDB. MongoDB Atlas integrates with key AWS services and combines operational data, metadata, and vector data into a single platform to let you quickly create dependable new AI experiences. Simplify your data administration, encourage extensive creativity, and deliver accurate user experiences backed by current business data.
MongoDB Atlas is secure by default. It utilises security components that are already in place throughout your deployment. Strong security measures protect your data in compliance with PCI DSS, GDPR, ISO 27001, HIPAA, and other regulations.
It is simpler to build complex RAG implementations when an operational database provides built-in vector search functionality. In reference to retrieval-augmented generation (RAG), an approach that leverages Large Language Models (LLM) enhanced by your own data to generate more accurate responses. Using MongoDB, you can store, index, and query vector embeddings of your data without the requirement for an additional add-on vector database.
Transform the way you design mobile apps with Atlas Device Sync. This completely managed device to cloud synchronisation solution helps speed up the development of superior mobile apps by your team.
Pricing for MongoDB Atlas
MongoDB Atlas offers a number of different pricing options based on your needs:
Is the MongoDB Atlas Free Tier Available? This tier is best suited for small production, testing, and development workloads. Every month, it can handle 50 writes, 100 million reads, and 512MB of storage.
MongoDB Atlas Cost Shared Cluster: This plan is ideal for one-off projects or low-traffic applications. It costs $9 per month.
Serverless: At $0.10 per million reads, this tier is a good option for apps whose traffic is erratic or unpredictable.
Dedicated Cluster: This tier offers the most control and scalability, starting at $57 per month (based on $0.08 per hour). It works well in production applications with high workloads.
Remember that these are only estimates, and the actual cost will vary based on your specific usage, which includes the services you choose, network traffic, storage requirements, and backup options.
Available right now
The MongoDB Atlas vector store in Knowledge Bases for Amazon Bedrock is accessible from both the US East (North Virginia) and US West (Oregon) regions. Make sure you view the whole Region list for planned updates.
News source: MongoDB Atlas AWS
0 Comments