MongoDB Atlas AWS
Billions of parameters are used in the training of foundational models (FMs) on massive amounts of data. However, they must refer to a reliable knowledge base that is distinct from the model’s training data sources in order to respond to inquiries from clients on domain-specific private data. Retrieval Augmented Generation (RAG) is a technique that is widely used to accomplish this. Without requiring the model to be retrained, RAG expands the capabilities of FMs to certain domains by retrieving data from the company’s own or internal sources. It is an affordable method of enhancing model output to ensure that it is accurate, relevant, and helpful in a variety of situations.
Without needing to create unique connections for data sources and manage data flows, Knowledge Bases for Amazon Bedrock is a fully managed feature that aids in the implementation of the complete RAG workflow, from ingestion to retrieval and quick augmentation.
The availability of MongoDB Atlas as a vector store in Knowledge Bases for Amazon Bedrock has been announced by AWS. You may create RAG solutions to safely link FMs in Amazon Bedrock to your company’s private data sources with MongoDB Atlas vector store connection. With this integration, the vector engines for Amazon OpenSearch Serverless, Pinecone, Redis Enterprise Cloud, and Amazon Aurora PostgreSQL-Compatible Edition are now supported by Knowledge Bases for Amazon Bedrock.
MongoDB Atlas and Amazon Bedrock Knowledge Bases build RAG apps
The vectorSearch index type in MongoDB Atlas powers vector search. The vector type of the field containing the vector data must be specified in the index specification. You must first establish an index, ingest source data, produce vector embeddings, and store them in a MongoDB Atlas collection before you can use MongoDB Atlas vector search in your application. In order to execute searches, you must first convert the input text into a vector embedding. After that, you may execute vector search queries against fields that are indexed as the vector type in a vectorSearch type index using an aggregation pipeline stage.
The majority of the labor-intensive work is handled by the MongoDB Atlas integration with Knowledge Bases for Amazon Bedrock. You can integrate RAG into your apps once the knowledge base and vector search index are set up. Your input (prompt) will be transformed into embeddings by Amazon Bedrock, which will then query the knowledge base, add contextual information from the search results to the FM prompt, and return the resulting response.
What is MongoDB Atlas
The most sophisticated cloud database service available, featuring built in automation for workload and resource optimization, unparalleled data distribution and mobility across AWS, Azure, and Google Cloud, and much more.
With a cloud database at its core, MongoDB Atlas is an integrated set of data services that streamlines and expedites the process of building with data. With a developer data platform that assists in resolving your data difficulties, you can build more quickly and intelligently.
Use apps anywhere
Use Atlas to run somewhere in the world. With AWS, Azure, and Google Cloud, you can deploy a database in over 90 regions and grow it to be global, multi-regional, or multi-cloud as needed. For extreme low latency and stringent compliance, pin data to specific areas.
Scale operations with assurance
Construct with assurance. Best practices are pre installed in Atlas, and it cleverly automates necessary tasks to guarantee that your data is safe and your database functions as it should.
Decrease the intricacy of the architecture
Using a single query API, access and query your data for any use case. The whole AWS platform, including full-text search, analytics, and visualisations, can instantaneously access data stored in Atlas.
Pay attention to the shipping features
Whether traffic triples or new features are added, don’t stop your apps from operating. In order to ensure that you always have the database resources you need to keep creating, Atlas includes sophisticated speed optimization capabilities.
MongoDB Atlas on AWS
Utilise AWS with MongoDB to create intelligent, enterprise ready applications. To assist you in rapidly developing reliable new AI experiences, MongoDB Atlas interacts with essential AWS services and unifies operational data, metadata, and vector data into a single platform. Streamline your data management, spur large scale innovation, and provide precise user experiences supported by up to date company data.
By default, MongoDB Atlas is secure. It makes use of security elements that are already present across your deployment. Robust security safeguards safeguard your data in accordance with HIPAA, GDPR, ISO 27001, PCI DSS, and other requirements.
Complex RAG implementations are easier to construct when an operational database has native vector search capabilities. Regarding retrieval-augmented generation (RAG), a technique that produces responses that are more accurate by utilizing Large Language Models (LLM) supplemented with your own data. You don’t need a separate add-on vector database in order to store, index, and query vector embeddings of your data using MongoDB.
Use Atlas Device Sync to transform your mobile app development process. With the help of this fully managed device to cloud synchronization solution, your team will be able to create better mobile apps more quickly.
MongoDB Atlas pricing
Depending on your requirements, MongoDB Atlas offers several different pricing tiers:
Is MongoDB Atlas Free
Free Tier: Small production, testing, and development workloads are best suited for this tier. It has 50 writes, 100 million reads, and 512MB of storage each month.
MongoDB Atlas Cost
Shared Cluster: This plan, which costs $9 per month, is best suited for individual projects or applications with little traffic.
Serverless: This tier is a suitable choice for apps with irregular or fluctuating traffic because it costs $0.10 per million reads.
Dedicated Cluster: This tier, which begins at $57 per month (calculated based on $0.08 per hour), provides the greatest control and scalability. Production applications with heavy workloads are best suited for it.
Keep in mind that these are only estimates; the real cost will vary depending on your unique usage, including the amount of storage needed, network traffic, backup choices, and other features.
Currently accessible
Both the US West (Oregon) and US East (North Virginia) regions offer access to the MongoDB Atlas vector store in Knowledge Bases for Amazon Bedrock. For upcoming updates, make sure to view the entire Region list.
