Side by Side with Elasticsearch & Solr part 2
summary
Benchmarks use JMeter hammering three m3.large-like VMs with Spanish Wikipedia, a 10M-video dataset and Apache logs; product search yields ~500 qps on Elasticsearch vs ~320 qps on Solr until Solr’s query-result cache skews results, and synthetic re-runs equalize indexes. Logs indexing grows from ~3.2K/s to ~4.6K/s when using time-based indices and hot/cold nodes and force-merge; Elasticsearch filters run about 2x faster, Solr does simple and nested aggregations faster (nested ~10x faster in these tests).
AI-written summary · about
chapters (10)
- 0:00Introduction and speakers
- 1:28Last year's recap and updates
- 5:24Theory behind horizontal scaling
- 8:59Product use case assumptions
- 11:05Product search performance results
- 15:05Logs use case overview
- 18:00Logs indexing performance
- 22:18Hot and cold node setups
- 27:17Hot and cold node results
- 29:52Summary and closing remarks
related talks
19mYou Know, for Search. Querying 24 Billion Records in 900ms
44mElasticsearch at Scale of Billions of Documents
43mCattle vs. Pets - A Cost-Effective Elasticsearch Architecture to Scale-Out...
43mMulti-tenancy for all workloads
35mHighly Available Search at Shopify
12mMilliSort and MilliQuery: Large-Scale Data-Intensive Computing in Milliseconds
38mLog Analytics: Understanding Complex Systems at Scale
39mScaling OpenSearch: Next-Generation Shard Allo...