Amagi’s intelligent media operations with Amazon Neptune

This article describes how Amagi built the Global Metadata Store (GMS) using Amazon Neptune and OpenSearch Service to manage complex media metadata across 2,500+ channels and 150+ countries.

• Neptune serves as the source of truth for RDF knowledge graphs, handling deep multi-hop semantic queries and rights inheritance
• OpenSearch Service provides a denormalized read cache for high-frequency transactional reads with sub-500ms latency
• Custom synchronization application resolves blank nodes (anonymous RDF nodes) from Neptune to searchable OpenSearch documents
• Hybrid query routing: Neptune handles deep lineage and rights queries; OpenSearch handles known-entity page loads and high-concurrency filters
• System manages 500M+ triples representing 1M+ media assets, sustaining 300–500 requests/second with peaks exceeding 1,000 requests/second
• RDF model enables schema-less metadata ingestion from studios and AI services without database migrations

The architecture demonstrates how combining graph databases with search indexes solves metadata complexity in media operations while maintaining performance at scale.