1 / 29

Queue Dynamics with Window Flow Control

Karl Henrik Johansson. H åkan Hjalmarsson. Steven Low. Kevin Tang. Lachlan Andrew. Krister Jacobsson – krister@caltech.edu. Queue Dynamics with Window Flow Control. TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: A A. Introduction. ACK. ACK.

adara
Download Presentation

Queue Dynamics with Window Flow Control

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Karl Henrik Johansson Håkan Hjalmarsson Steven Low Kevin Tang Lachlan Andrew Krister Jacobsson – krister@caltech.edu Queue Dynamics with Window Flow Control TexPoint fonts used in EMF. Read the TexPoint manual before you delete this box.: AA

  2. Introduction ACK ACK Window = Outstanding Packets 1101 1001 1101 1101 1001 0101 0101 1101

  3. TCP is window based • TCP carries >80% of the Internet traffic • TCP limitations • Bandwidth scalability issues • Wireless links • Delay dependent resource allocation • New designs needed! Window based? • Tractable flow-level models valuable Background

  4. Packet-level Internet is extremely complex Abstract away packet level detail Model flows of packets as fluids Feedback mechanism ODEs Powerful analysis frameworks Used to reverse engineer TCP Fluid Flow Modeling

  5. Motivation

  6. Window = Outstanding Packets WindoW Based Congestion Control ACK 1101 1101

  7. WindoW Based Congestion Control Round Trip Time (RTT)

  8. WindoW Based Congestion Control Sending rate is dependent on the network state! Sending rate per RTT = Window 1101 1101

  9. Window Based Congestion Control • Routers operate buffers • Queuing delay • RTT = Propagation delay + Queuing delay • Control system • ACK-Clocking

  10. Window Based Congestion Control

  11. Input: Window size • Output: Queue size • Assumptions • No loss • FIFO buffering • Single link • No forward propagation delay • Operate far from static nonlinearities Modeling

  12. Modeling a queue • Integrates the link excess rate • Rate of change:  Queue:

  13. What is xn(t)??? Common approximation: xn(t)¼wn(t)/¿n(t) Does not consider ACK-clocking! Modeling the Instantaneous Rate

  14. Modeling the Instantaneous Rate 1101 1101 t0

  15. Queue integration: Instantaneous rates: Round trip time = Prop. Delay + Queuing Delay: Model Summary

  16. Step response, single source, single bottleneck Model Validation UDP cross traffic!

  17. Analyzing the effect of cross traffic

  18. Analyzing the stepwise convergence

  19. More Model Validation

  20. Unique equilibrium Rates xn(t) non-unique Equilibrium

  21. Single bottleneck locally stable from windows to queue Not stable from windows to rates when flows’ RTT ratios are rational Stability Rate (FFT) Queue (FFT) Queue (FFT) Rate (FFT)

  22. Multilink networks may be unstable from windows to the queues! Stability

  23. Previous models appears as approximations of the integral • Numerical quadrature • Taylor approximations, Padé approximations • Low order models valid for small RTTs Approximations and Previous Work

  24. Closing the Loop Queuing Delay FAST TCP

  25. Previous models: locally asymptotically stable • Proposed model: system destabilizes for • Heterogeneous RTTs • Rational RTT ratios • NS simulations and testbed experiments confirms predictions Case study: FAST TcP

  26. FAST TCP designed to be stable for all delays • Scale down gain inversely proportional to RTT • Feedback on the scale of other flows’ RTT • Need to scale down gain inversely proportional to other flows’ RTTs • Alternatively: attenuate the large delay feedback • Gain is dependent on sending rate • Punish large delay flows • TCP Reno: RTT biased resource allocation!!! Some remarks

  27. New accurate model of the ACK-clocking mechanism Unstable for certain configurations!!! Dynamics more complex than previously known May have impact on existing (stability) results Model validation is crucial Equally fair window based congestion control problematic Window based or rate based congestion control? Conclusions

  28. WindoW Based Congestion Control Window size = 2 Window size = 1 1101 1101

  29. Window Based Congestion Control Modeling Model Properties Model Application Conclusions Outline

More Related