Maximize media search scalability & performance w/
summary
Sony Liv migrates from ElasticSearch to OpenSearch 2.11→2.13 and runs a 45-data-node + 3-master + 3-coordinator cluster indexing 500M+ docs, serving 13k hits/s with median search latency around 4 ms after a zero-downtime migration. Recommendations use OpenSearch KNN/vector features plus user/item collaborative filters and MLP models. Plans include moving CMS, DFS and recommender indexes into OpenSearch, using AWS-managed deployment for operability, cross-cluster search for microservice domains, and relying on OpenSearch ML and security roadmap to close remaining gaps while cutting operational cost.
AI-written summary · about
chapters (9)
- 0:00SonyLIV search intro
- 1:11Migrating to OpenSearch
- 2:59OpenSearch scale metrics
- 4:03Zero-downtime migration gains
- 5:28Key reasons for choosing OpenSearch
- 8:01Cross-cluster search & privatization
- 9:26Security trade-offs & roadmap
- 10:31Search & recommendation integration
- 11:36Conclusion
related talks
40mBoost your workloads with OpenSearch innovations
47mWhy, What, When and How of Running Diff Types of Sear...
48mUnlocking Next-gen Semantic Search With OpenSearch 3.x
36mSearch Complete
26mOpenSearch and the Path to a Trillion-s...
40mAgentic Relevance Tuning With OpenSearch: Faster, Easier, and Higher Quality Search at Scale
05mKeynote: OpenSearch 3.0 and Lucene 10, Paving the Way for AI and Vector Innovation
31mIntelligent Japanese Search With OpenSearch