Optimizing Cost and Performance with Arm64

Optimizing Cost and Performance with Arm64

USENIX via YouTube Direct link

Migrated prod Shepherd

21 of 30

21 of 30

Migrated prod Shepherd

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

Optimizing Cost and Performance with Arm64

Automatically move to the next video in the Classroom when playback concludes

  1. 1 Intro
  2. 2 WTF is architecture? Why multiarch?
  3. 3 History: 80s, 90s, 00s, 10s, and beyond
  4. 4 If it ain't broke...
  5. 5 ARM is more efficient.
  6. 6 Data storage engine and analytics tool
  7. 7 Service Level Objectives (SLO)
  8. 8 SLOs are user flows
  9. 9 Same reliability, lower costs with ARM6
  10. 10 Complexity stayed manageable
  11. 11 Prod: customers observe data
  12. 12 Kibble observes dogfood
  13. 13 Dogfood observes prod
  14. 14 Service Architecture
  15. 15 Shepherd: ingest API service
  16. 16 Is it feasible to migrate?
  17. 17 Producing artifacts for Arm64
  18. 18 Initial findings
  19. 19 A/B testing
  20. 20 Dogfood Shepherd cost reduction
  21. 21 Migrated prod Shepherd
  22. 22 Migrated prod Retriever
  23. 23 AWS ran out of m6gd spot instances
  24. 24 Kafka + the long tail
  25. 25 Graviton2 going strong
  26. 26 Have a measurable goal in mind
  27. 27 Acknowledge hidden risks
  28. 28 Take care of your people
  29. 29 Optimize for safety
  30. 30 Graviton2 blog posts

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.