Orange Stream
summary
Three short case studies with concrete takeaways: a serverless benchmark shows whole-app AWS costs make Lambdas only about 10–15% of monthly spend, so swapping languages often saves little; a real JSON→CSV pipeline comparison finds GraalVM native images sometimes run faster but suffer a surprising S3 GET slowdown — Graal downloads twice as slow for some workloads — and overall cost can be higher. Azure Functions with Java reports frequent cold starts and awkward local debug. A Helidon demo claims Nema (virtual‑thread) server matches reactive throughput with blocking code.
AI-written summary · about
chapters (6)
- 0:00"Serverless 2049" - Mykyta Berehulia
- 1:06:00Break
- 1:14:56"Do Not Use Azure Functions with Java!" - Mikhail Kotor
- 2:05:55Break
- 2:30:00"Levels of Docs as Code"- Peter Verhas
- 3:32:00"Reactive vs. Virtual threads power with Helidon Nima" - Oleg Klimenko
related talks
48mJava in the Clouds: Pointless and Expensive or Productive and Must-Have?
51mAdapting Java for the Serverless world!
36mServerless Java
32mServerless & JVM: Build Your Benchmark to Bust Preconceptions
01h 00mAdapting Java for the Serverless world
53mHigh performance Serverless Java on AWS
56mThinking in Serverless Java (Code)
01h 03mKeynote: Why The Future of Java in the Cloud is Serverless
