OpenSearch & Vector Search

Open source search and vector databases at any scale.

5 sessions at the summit4 external resources

Overview

Amazon OpenSearch Service is a managed search and analytics engine forked from Elasticsearch. It powers full-text search, log analytics (the ELK pattern), and — importantly for modern AI — vector search for semantic retrieval and RAG. OpenSearch Serverless eliminates capacity management, while OpenSearch Ingestion provides a managed data pipeline. The OR1 storage option delivers up to 11x better indexing throughput.

Key concepts

  1. Inverted index (lexical) vs. HNSW (vector) search
  2. Hybrid search — combining BM25 and vector scores
  3. Index lifecycle management and tiered storage
  4. k-NN plugin and quantization for memory efficiency
  5. Observability: ingest pipelines, dashboards, alerting

Key AWS services

  • Amazon OpenSearch Service
  • OpenSearch Serverless
  • OpenSearch Ingestion

Learn more — curated resources

Hand-picked official docs, foundational papers, and the best community guides for going deeper on this topic.

Sessions on this topic

5 sessions from the Summit covered this topic. Each is a self-contained mini-lesson.

  1. ANT301Advanced

    A practitioners guide to data for agentic AI

    In this session, gain the skills needed to deploy end-to-end agentic AI applications using your most valuable data. This session focuses on data management using processes like Model Context Protocol (MCP) and Retrieval Augmented Generation (RAG), and provides concepts that apply to other methods of customizing agentic AI applications. Discover best practice architectures using AWS database services like Amazon Aurora and OpenSearch Service, along with analytical, data processing and streaming experiences found in SageMaker Unified Studio. Learn data lake, governance, and data quality concepts and how Amazon Bedrock AgentCore and Bedrock Knowledge Bases, and other features tie solution components together.

  2. STP205Intermediate

    How Dovetail powers Multi-Tenant Agents with Vector Indexing at Scale

    When you're building multi-tenant vector search but can't control what customers throw at you, every indexing decision — isolation, embedding, chunking, partitioning — becomes a bet you're making blind, and this talk gives you the framework to make the right ones.

  3. PRT112-SFoundational

    Empower Data with Oracle AI Database and Native AI Services on AWS

    Organisations today depend on fast, secure access to data to support mission-critical operations and evolving AI workloads. With Oracle AI Database services available on AWS, your team can streamline data integration and deliver high-impact solutions for semantic search, fraud detection, quality control, and much more.

  4. DAT402Expert

    Deep dive into database integrations with AWS Zero-ETL

    Learn how AWS zero-ETL integrations eliminate complex data movement pipelines across multiple database engines, enabling data engineers, architects, and DBAs to reduce maintenance overhead while ensuring near real-time data availability for analytics and ML workloads. Examine the underlying architecture for supported zero-ETL integrations between Amazon Aurora, Amazon DynamoDB, and Amazon RDS sources to Amazon Redshift, Amazon SageMaker, and Amazon OpenSearch Service targets. Explore data movement options, tunable settings, and monitoring capabilities for ongoing data replicationall without traditional ETL complexity.

  5. MAE205Intermediate

    AI at Speed of News: Unlocking Value from Media with Generative AI

    For media and communications organizations, the ability to rapidly discover, repurpose, and distribute content across platforms directly impacts revenue and audience engagement. This session examines how Generative AI is transforming content operations through intelligent metadata extraction, semantic search, and automated workflow orchestration. Using a case study from a global media organization managing 13 petabytes of content growing at 3,000 hours daily, we'll explore practical implementations using Amazon OpenSearch for multimodal retrieval, Amazon Neptune for knowledge graphs, and agentic AI for content assembly. Learn how organizations are achieving faster time-to-market, improved content monetization, and enhanced audience experiences through AI-powered content discovery and recommendation systems

Live updates related to this topic LIVE

Sourced via Parallel AI Monitor — continuous web watch on 21 topical streams. Updated .

External links matched to this topic via topic relevance. The KB does not endorse third-party content; verify before citing.

Non-obvious insights

From the Playbook

One sharp, contrarian insight per session — the things teams don't think of unprompted.

RAG retrieval quality is dominated by chunking strategy, not embedding model. Boring but true. Spend a week on chunk size, overlap, and semantic boundaries before you spend a dollar on a fancier embedder. ---ANT301 — A practitioners guide to data for agentic AI
Most teams optimise for retrieval *quality* and forget *tail latency*. A single cold tenant with a giant document set will kill P99 for everyone unless you isolate aggressively. Per-tenant query budgets are a feature, not a limitation. ---STP205 — How Dovetail powers Multi-Tenant Agents with Vector …
Putting vectors next to transactional data unlocks real-time RAG over fresh data. Your warehouse can't do that — there's always replication lag. If your AI use case requires acting on fresh transactional data (fraud detection, real-time personalisation), the consolidated DB option becomes more compelling than its raw benchmarks suggest. ---PRT112-S — Empower Data with Oracle AI Database and Native AI S…
Multimodal retrieval (vision + text + audio together) is finally good enough for production media use cases. The unlock is *search-by-vibe* ("find me clips like this one"), not just keywords. That changes editorial workflows fundamentally — editors become curators, not searchers. ---MAE205 — AI at Speed of News: Unlocking Value from Media with…