Reindexer - a very fast in-memory DB with full-text search
summary
Reindexer is presented as a very fast in-memory database with built-in full-text search used to power an IPTV/OTT platform serving about 10 million users. The talk explains using Reindexer to satisfy high-load backend requirements: delivering prefiltered, sorted data slices for thin clients according to complex business logic. It focuses on performance and API behavior under heavy load for large-scale content delivery.
AI-written summary · about
chapters (11)
- 0:00About Restream
- 1:11What is Reindexer
- 2:15New platform challenge
- 11:50Building Reindexer
- 15:01How Reindexer works
- 18:36Operation modes
- 21:06Connection interfaces
- 22:10Performance comparison
- 31:25Why Reindexer is fast
- 34:20Future plans
- 41:30Q&A
related talks
28mEnhancing Elasticsearch Performance: Innovative Reindexing...
41mAn introduction to search-index.js
36mSearch on the fly: how to lighten your Big Data
28mSearch and Sushi; Freshness Counts
53mIngest node: (re)index and enrich documents in Elasticsearch
22mHypeReca: Distributed Heterogeneous In-Memory Embedding Database for Training...
12mREMIX: Efficient Range Query for LSM-trees
14m5 great ways to reimagine your data layer