TCP LoLa Toward Low Latency and High Throughput Congestion Control - - PowerPoint PPT Presentation

▶
tcp lola toward low latency and high throughput
SMART_READER_LITE
LIVE PREVIEW

TCP LoLa Toward Low Latency and High Throughput Congestion Control - - PowerPoint PPT Presentation

Jul 09, 2023 408 likes •577 views

TCP LoLa Toward Low Latency and High Throughput Congestion Control Mario Hock, Felix Neumeister, Martina Zitterbart, Roland Bless Institute of Telematics (TM) KIT The Research University in the Helmholtz Association www.kit.edu

slide-1
SLIDE 1

KIT – The Research University in the Helmholtz Association

Institute of Telematics (TM)

www.kit.edu

TCP LoLa – Toward Low Latency and High Throughput Congestion Control

Mario Hock, Felix Neumeister, Martina Zitterbart, Roland Bless

slide-2
SLIDE 2

2 Institute of Telematics (TM) http://tm.kit.edu/

Motivation High Throughput and Low Delay

Typically considered as conflicting goals or trade-off Not necessarily so: mitigate this trade-off Approaches: AQM, Tweaks to existing CC (e.g., Alternative Back-

  • ff with ECN), New congestion controls

Investigate how far we can get with a congestion control

Low queuing delay High utilization/throughput Scalable (also 10 Gbit/s and beyond) RTT Fairness Should work with regular tail-drop queues Focus: Wide area networks (not Data Center)

IETF 100 Singapore - ICCRG

slide-3
SLIDE 3

3 Institute of Telematics (TM) http://tm.kit.edu/

Objective and Challenges General goal

Determine a suitable amount of inflight data

achieving high bottleneck link utilization avoid creating standing queues  keep queuing delay low

Configurable fixed target delay value

Congestion: persistent queuing delay above fixed target

Challenge

Convergence to fairness: total amount of inflight data okay, but maybe unequal rate shares Increase inflight data of one sender while reducing it for others Interaction with small queue is more difficult Without sacrificing the low delay goal!

IETF 100 Singapore - ICCRG

slide-4
SLIDE 4

4 Institute of Telematics (TM) http://tm.kit.edu/

TCP LoLa (Low Latency) Scalable approach based on queuing delay thresholds

e.g., sender sets target= 5 ms

Estimate queuing delay

using min filter over fixed time period  measure standing queue heuristic to adapt to network path changes (e.g., increasing )

Congestion window-based approach

packet pacing is beneficial, but not necessary

IETF 100 Singapore - ICCRG

Bottleneck buffer in

  • ut

target low

Congestion High throughput, low delay Link utilization unknown

slide-5
SLIDE 5

5 Institute of Telematics (TM) http://tm.kit.edu/

TCP LoLa (Low Latency) Flow States

IETF 100 Singapore - ICCRG

Novel mechanism: See next slide Reliably detect congestion event + Synchronization Drain queue completely, Keep utilization high

target low

Congestion High throughput, low delay Link utilization unknown

Slow Start Cubic Increase Fair Flow Balancing CWnd Hold max min > 2low

  • delay target
  • delay low

[fixed waiting time sync]/ Action: Tailored Decrease

slide-6
SLIDE 6

6 Institute of Telematics (TM) http://tm.kit.edu/

Fair Flow Balancing Novel convergence to fairness mechanism

Equalize amount of data that each flow may queue at the bottleneck Dynamically scale allowed amount of data w.r.t. given delay target Knowledge about current shares not available

Key concepts

Flow with more than fair share keeps CWnd Flow with smaller than fair share increases CWnd Allowed amount of data in queue (fair queue share) is time dependent: Xt

  • , : time since target exceeded, : constant

IETF 100 Singapore - ICCRG

slide-7
SLIDE 7

7 Institute of Telematics (TM) http://tm.kit.edu/

Testbed Setup

Implemented as Linux Kernel module (Ubuntu 16.04) Packet pacing enabled Traffic generated with iperf3 low= 1ms, target= 5ms, sync= 250ms, measure= 40ms

IETF 100 Singapore - ICCRG

WAN link 21ms RTT

slide-8
SLIDE 8

8 Institute of Telematics (TM) http://tm.kit.edu/

Evaluation Results – Throughput/Fairness

High Throughput Convergence to Fairness Link cannot be fully utilized Misinterprets jitter as congestion

IETF 100 Singapore - ICCRG

TCP LoLa TCP Vegas

slide-9
SLIDE 9

9 Institute of Telematics (TM) http://tm.kit.edu/

Low (Queuing) Delay

Delay kept around target (~5ms queuing delay) Fills buffer completely

IETF 100 Singapore - ICCRG

TCP LoLa CUBIC TCP

slide-10
SLIDE 10

10 Institute of Telematics (TM) http://tm.kit.edu/

Queuing Delay – RTT Independent

IETF 100 Singapore - ICCRG

Single flow, base RTT varied: 5ms, 61ms, 101ms LoLa keeps delay around target (~5ms queuing delay) Queuing delay is independent of

Base RTT (and rate – not shown here) Number of senders (next slide)

slide-11
SLIDE 11

11 Institute of Telematics (TM) http://tm.kit.edu/

Several Flows Starting in Succession

RTTs measured at TCP senders LoLa still controls overall queuing delay No packet loss Starting flows fill the buffer Delay kept around 2 > 1 Mio retransmissions

IETF 100 Singapore - ICCRG

BBR TCP LoLa

slide-12
SLIDE 12

12 Institute of Telematics (TM) http://tm.kit.edu/

RTT Fairness

Flow 1: 21 ms base RTT Flow 2: 101 ms base RTT Small buffer: 12.5 Mbyte Convergence to fair shares RTT unfairness Convergence to similar congestion windows No convergence to fair rate shares

IETF 100 Singapore - ICCRG

TCP LoLa CUBIC TCP

slide-13
SLIDE 13

13 Institute of Telematics (TM) http://tm.kit.edu/

Conclusion and Outlook First tests of the overall concepts [1]

Parameters not thoroughly optimized Not a full-fledged TCP variant yet

Further investigations

Use of one-way delay instead of RTT measurements Influence of delayed and compressed ACKs Performance in wireless environments Multiple bottleneck scenarios Coexistence with loss-based variants (separate queues, AQMs, …) [2]

Planned

Use of explicit feedback from the network

IETF 100 Singapore - ICCRG

slide-14
SLIDE 14

14 Institute of Telematics (TM) http://tm.kit.edu/

References

[1] M. Hock, F. Neumeister, M. Zitterbart, R. Bless: „TCP LoLa: Congestion Control for Low Latencies and High Throughput“, IEEE 42nd Conference on Local Computer Networks, Singapore, Oct 9–12, 2017 [2] M. Hock, R. Bless, M. Zitterbart, „Toward Coexistence of Different Congestion Control Mechanisms”, 2016 IEEE 41st Conference on Local Computer Networks, pp. 567–570, Dubai, United Arab Emirates, November 2016

IETF 100 Singapore - ICCRG

slide-15
SLIDE 15

15 Institute of Telematics (TM) http://tm.kit.edu/

BACKUP SLIDES

IETF 100 - ICCRG, Singapore

slide-16
SLIDE 16

16 Institute of Telematics (TM) http://tm.kit.edu/

Formulas

RTT measurements

  • minRTT | ∈ measure,
  • ≔ min
  • ,
  • Reset
  • if
  • hasn‘t been measured

close

  • for 100 tailored decreases
  • delay=
  • Cubic Increase / Tailored Decrease

CWnd · CWndmax CWndmax

  • · CWndmax
  • · /
  • Fair Flow Balancing

Xt

  • , : time since target exceeded,

X allowed amount of queued data, : constant

IETF 100 Singapore - ICCRG