Adapting Java for the Serverless world!

on YouTube ↗51m

" summary

Java on AWS faces high cold starts and memory costs but practical fixes exist: using smaller bundles and async AWS SDK cuts download and init time; choosing more memory/ARM and tuning with Power Tuning changes cost-performance (example: 1M requests at 512MB and 200ms yields ~$0.20 request + $1.67 duration); GraalVM native image gives sub‑1s cold starts but requires Linux custom runtimes, 6–10GB build RAM and longer CI; AWS SnapStart (CRaC) snapshots runtime during deploy to restore fast (deploy ~2.5m, x86 only, no EFS/provisioned concurrency); Micronaut/Quarkus/Spring (AOT support) reduce reflection and work with these approaches.

AI-written summary · about

" chapters (59)

1. 0:00Intro
2. 0:46About Vadym
3. 1:35About Iplabs
4. 2:47Java popularity
5. 3:12AWS
6. 3:30Challenges
7. 5:42Java Community
8. 6:00Serverless Adoption
9. 6:29Why Serverless
10. 6:42What are these things
11. 6:53Two call start times
12. 7:03Lambda function
13. 7:47Java Functions
14. 8:57Price Model
15. 9:19Duration
16. 9:42Parameter
17. 9:52Example
18. 10:18Basics
19. 11:40How it works
20. 13:56Serverless Architecture
21. 14:38Cold Start Times
22. 15:56First Function
23. 17:10Credentials
24. 18:02HTTP Client
25. 21:37Time
26. 23:31TAST Compilation
27. 24:53General Best Practices
28. 26:12Memory Configuration
29. 26:34Output
30. 26:57Arm
31. 27:25AWS Compute Optimizer
32. 28:10AWS Cold Start
33. 28:40substrate VM
34. 28:50ahead of time compilation
35. 29:16native image
36. 30:04custom runtime
37. 30:15deploy on AWS
38. 30:40guten VM
39. 31:13productivity
40. 32:34principles
41. 33:26steps to deploy
42. 34:00Micronaut
43. 34:14Spring
44. 34:44No AWS
45. 36:06Cold Stone
46. 36:35GroundVM
47. 37:14Conclusions
48. 37:24Challenges and tradeoffs
49. 38:46Snapstart
50. 40:13JVM Snapshot
51. 42:17JVM Snapshot Statistics
52. 42:43AWS Snap Snapshot
53. 43:24Snap Snapshot ReStore
54. 46:07Framework Priming
55. 47:03Open Source Project
56. 47:43Engineering Problem
57. 48:07Latency
58. 48:48Snapstart Limitations
59. 49:45Conclusion

" related talks

• 01h 00mAdapting Java for the Serverless worldVadym Kazulkin · Voxxed Days Luxembourg 2022
• 51mHow To Reduce Cold Starts for Java Serverless Applications in AWSVadym Kazulkin · GOTO Amsterdam 2024
• 53mHigh performance Serverless Java on AWSVadym Kazulkin · GeeCON 2024
• 46mReduce Cold Starts for Java Serverless Applications in AWSVadym Kazulkin · JCON WORLD ONLINE 2023
• 56mThinking in Serverless Java (Code)Adam Bien · GeeCON Prague 2023
• 53mAdapting Java for the Serverless worldVadym Kazulkin · JCON-ONLINE 2021
• 19mTen things that improve the performance of your Java based AWS LambdaJeroen Reijn · J-Fall 2022
• 56mJava on AWS - all the details that matterSebastian Gebski · GeeCON 2024