Overload Control for µs-scale RPCs with Breakwater

Overload Control for µs-scale RPCs with Breakwater

USENIX via YouTube Direct link

Intro

1 of 24

1 of 24

Intro

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

Overload Control for µs-scale RPCs with Breakwater

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

  1. 1 Intro
  2. 2 Trend: High Fan-out Internet
  3. 3 Causes of Server Overload
  4. 4 Performance Without Overload Control
  5. 5 Ideal Overload Control
  6. 6 Strawman #1: Server-side AQM
  7. 7 Strawman #2: Client Rate limiting
  8. 8 Breakwater Overload control scheme for us-scale RPCs
  9. 9 Breakwater's benefits Handles server overload with us-scale RPCs with
  10. 10 Queueing delay as congestion signal
  11. 11 Credit-based admission control Breakwater controls amount of incoming requests with credits
  12. 12 Demand Message Overhead Server needs to know which client has demand
  13. 13 Impact of Credit-based Admission Control Credit-based admission control has lower and bounded tail latency but lower throughput.
  14. 14 Piggybacking Demand Information Breakwater piggybacks clients' demand information into requests.
  15. 15 Demand Speculation Breakwater speculate clients' demand to minimize message overhead
  16. 16 Impact of Adding Demand Speculation Demand speculation improves throughput with higher tail latency
  17. 17 Credit Overcommitment Server issues more credit than the number of requests it can accomodate
  18. 18 Incast Causing Long Queue With credit overcommitment, multiple requests may arrive at the server at the same time
  19. 19 Delay-based AQM To ensure low tail latency, the server drops requests if queueing delay exceeds threshold.
  20. 20 Impact of Adding Delay-based AQM Breakwater achieves high throughput and low and bounded tail latency at the same time
  21. 21 Evaluation
  22. 22 High Goodput with Fast Convergence
  23. 23 Fast Notification of Reject
  24. 24 Conclusion • Breakwater is a server-driven credit-based overload

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.