NoSQL, Data Processing, and Business Transparency with RavenDB
summary
FinStat uses RavenDB as its primary store since 2013 to power FinStat.sk and HitHorizons.com, holding tens of millions of documents (hundreds of GB), serving hundreds of thousands of users monthly and millions of pages. Data pipelines use pre-processing plus map-reduce indexes for transformation and fast queries (low hundreds of ms across tens of millions). Production runs a three-node cluster on local NVMe with cloud backups; lessons include avoiding giant documents, spreading import load, and validating new DB versions after a 5x performance jump on v4.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker setup
- 2:09About the speaker and RavenDB
- 3:36Company overview and projects
- 10:23Why RavenDB was chosen
- 13:16RavenDB features and benefits
- 18:47Cloud infrastructure and performance
- 22:19High availability and backups
- 25:13Challenges and pitfalls with RavenDB
- 30:16Support experience and upgrades
- 31:41Performance improvements over time
- 32:48When to use RavenDB
- 35:20Conclusion and competitive edge
- 37:27Q&A session
related talks
46mHighly Available & Scalable Solutions with RavenDB
30mModernizing data collection at FINRA with Amazon DocumentDB
31mFrom Raw Data to Real Impact: Lessons from Real-World Transformations
01h 03mExtreme Performance Architecture
52mData storage for modern applications - Mike Lehan
27mStorage and analytics
55mData Modeling in a Non-Relational World
01h 03mRavenDB: Practical Considerations for ACID/MVCC Storage Engines