Capacity Planning, and Scaling/Optimization for Vector Workloads

Capacity Planning, and Scaling/Optimization for Vector Workloads - Jon Handler, Amazon Web Services With the advent of AI-powered search, OpenSearch’s vector database has become a key component for more accurate search, and for ChatBots and AI Agents. If you want to adopt AI-powered search, you need to determine a cluster size that will support your workload at your target performance, accuracy, and cost. OpenSearch’s diverse algorithm, engine, and quantization options make predicting cluster size accurately an even more complicated proposition. If you’re already underway, and revisiting your sizing, you need to understand the key tradeoffs that will enable you to reduce cost, improve performance, and hit your accuracy target. This session will help you find the answers! You’ll learn about the key elements of sizing OpenSearch clusters, the different algorithms and their tradeoffs, and best practices for building a durable, correctly-scaled OpenSearch cluster.