Building .NET software with focus on performance and scalability
summary
Warren builds a .NET investment core that shifts from MSSQL to Amazon Aurora MySQL with up to 15 read replicas to serve 92% of requests as reads (99.5% in practice). It uses Entity Framework for simple CRUD, Dapper for hot queries and stored procedures for batched transforms; one portfolio query drops from ~13s to 0.4s (≈32.5× faster) after removing EF context, and a nightly job falls from ~6 hours to ~8 minutes (≈45× faster). Web APIs run self-hosted (no IIS) and processing uses prioritized queues and in‑memory caches.
AI-written summary · about
chapters (12)
- 0:00Intro
- 3:45System overview
- 7:25Performance
- 16:52What we lost
- 18:30Code maintenance
- 22:24Database access
- 24:02Results
- 26:15Complex process
- 27:57Performance is essential
- 28:28Microservices
- 30:23Microwaves
- 30:53Summary
related talks
12mWhen each millisecond counts?
47m.NET Blub: Frameworks beyond Microsoft
01h 03mExtreme Performance Architecture
48mWhy I’m Not Leaving .NET
50mHow I have improved the Database performance from 5 seconds to 1 millisecond
48mAnalytics Architectures on Microsoft Fabric
37mRock Your Code Code and App Performance for Microsoft .Net
01h 02mWriting high performance code in .NET