.NET Rocks Panel Discussion: Going Serverless
summary
Panelists compare AWS Lambda, Azure Functions and Firebase with concrete tradeoffs: 3 second cold start and ~30 minute idle decay on Azure functions, a real project using the AWS free tier serving 20–30k users, and one team managing 140 Lambdas. They show platform patterns (S3/DynamoDB/Aurora, BigQuery, Cosmos DB), pitfalls like recursive event loops that spam email, observability via X‑Ray/Stackdriver/App Insights, and a BigQuery demo reducing 82M rows to 2.5k in under 30s for $5.
AI-written summary · about
chapters (23)
- 0:00Welcome to Going Serverless panel
- 2:12What is serverless
- 4:18Serverless with AWS Lambda
- 5:42Serverless vs containers & microservices
- 8:51When to use serverless
- 9:32Firebase & real-time triggers
- 11:17Hybrid architectures & monoliths
- 13:02Teams skipping containers
- 14:04API management & bad actors
- 15:49DevOps for serverless
- 16:31Portability & lock-in risks
- 18:59Break & sponsors
- 27:21Giveaway & sponsor segment
- 29:07Panelists' tech wish lists
- 31:35PWA & serverless future
- 34:01Serverless data & ML services
- 36:47What is serverless data
- 38:13Machine learning in serverless architectures
- 39:16Cost & performance
- 41:45Cosmos DB scaling
- 42:47Cold starts & warming
- 44:12What panelists build next
- 45:39Closing remarks
related talks
47mAdvanced Serverless Architectural Patterns on AWS
37mPerformance Tales of Serverless
01h 02mExploring Serverless Options for .NET in Azure, AWS, and Beyond
54mServerless beyond Functions
30mPerformance Tales of Serverless
35mConvergence of Serverless APIs & Compute
42mNo Server Is Better Than No Server: Building Blocks for Serverless Architectures
01h 00mNo, I don't need servers for .NET development in AWS!