TCP-Friendliness of SCTP and Concurrent Multipath Transfer (CMT) - - PowerPoint PPT Presentation

▶
tcp friendliness of sctp and concurrent multipath
SMART_READER_LITE
LIVE PREVIEW

TCP-Friendliness of SCTP and Concurrent Multipath Transfer (CMT) - - PowerPoint PPT Presentation

Feb 23, 2023 526 likes •823 views

TCP-Friendliness of SCTP and Concurrent Multipath Transfer (CMT) ILKNUR AYDIN ICCRG meeting @PFLDNET Nov 28-29 2010, Lancaster, PA, USA Outline Background on TCP-Friendliness (TCP-F) TCP-F of single-homed SCTP Motivation, SCTP vs. TCP

slide-1
SLIDE 1

ILKNUR AYDIN

ICCRG meeting @PFLDNET

Nov 28-29 2010, Lancaster, PA, USA

TCP-Friendliness of SCTP and Concurrent Multipath Transfer (CMT)

slide-2
SLIDE 2

Outline

  • Background on TCP-Friendliness (TCP-F)
  • TCP-F of single-homed SCTP
  • Motivation, SCTP vs. TCP mechanics
  • Experimental Framework
  • Results and Analysis
  • Conclusions
  • TCP-F of SCTP-based CMT
  • Motivation
  • Experimental Framework
  • Results and Analysis
  • Discussion and Conclusions
slide-3
SLIDE 3

Background: TCP-Friendliness (TCP-F)

3

slide-4
SLIDE 4

Background: TCP-Friendliness (TCP-F)

4

slide-5
SLIDE 5

TCP-Friendliness (TCP-F)

5

  • “definition”: a non-TCP flow should not

consume more resources than a confirming TCP flow under the same conditions + implement some form of congestion control mechanism

* by Mahdavi and Floyd (1997), revised by Padhye (1998) and others later on

Bandwidth consumed by a TCP flow* = 1.22 x MSS / RTT x sqrt(loss)

slide-6
SLIDE 6

6

  • Comprehensive SCTP simulation module for

the QualNet simulator

SCTP QualNet Module

slide-7
SLIDE 7

TCP-Friendliness of single-homed SCTP

slide-8
SLIDE 8

SCTP vs. TCP Mechanics

8

  • Transport Protocol Overheads
  • Transport PDU headers
  • Message-based (SCTP) vs. byte-based transmission (TCP)
  • Transport ACKs
  • Congestion Control Mechanisms
  • SCTP is “similar” to TCP but already has some of the TCP

enhancements (SACK, ABC, initial cwnd size, … )

Hypothesis: SCTP throughput may be better than TCP’s under the same conditions.

slide-9
SLIDE 9

Experimental Framework

9

5, 10, or 20 Mbps 45 msec

  • Flow: a greedy application over SCTP or TCP (flow 1 from

S1 to D1, flow 2 from S2 to D2

  • Case-I: Two flows start at the same time (how two flows

grow together?)

  • Case-II: Latter flow starts after the earlier is at steady-state

(how one flow gives way to another flow?)

  • Metrics: Throughput, Transport Load, Goodput, Fairness

Index, Link Utilization, System Utilization

slide-10
SLIDE 10

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 20 10 5 Normalized Throughput Tight Link Bandwidth (Mbps) flow 1 flow 2

Flows Start at the Same Time

10

SCTP TCP-SACK

slide-11
SLIDE 11

100 200 300 400 500 0.5 1 1.5 Simulation Time (sec) Normalized Throughput

TCP-SACK and SCTP flows grow together

11

SCTP TCP-SACK

slide-12
SLIDE 12

60 65 70 75 80 0.5 1 1.5 Simulation Time (sec) Normalized Throughput

TCP-SACK and SCTP flows grow together

12

SCTP TCP-SACK

slide-13
SLIDE 13

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 20 10 5 Normalized Throughput Tight Link Bandwidth (Mbps) flow 2 flow 11 flow 12 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 20 10 5 Normalized Throughput Tight Link Bandwidth (Mbps) flow 2 flow 11 flow 12

One Flow gives way to another Flow

13

TCP-SACK gives way to SCTP

SCTP TCPS1 TCPS2

SCTP gives way to TCP-SACK

TCPS SCTP2 SCTP1

slide-14
SLIDE 14

200 400 600 800 1000 1200 1400 1600 0.5 1 1.5 2 Simulation Time (sec) Normalized Throughput

SCTP gives way to TCP-SACK

14

SCTP TCP-SACK

slide-15
SLIDE 15

75 80 85 90 95 100 105 0.5 1 1.5 2 Simulation Time (sec) Normalized Throughput

SCTP gives way to TCP-SACK

15

SCTP TCP-SACK

slide-16
SLIDE 16

200 400 600 800 1000 1200 1400 1600 0.5 1 1.5 2 Simulation Time (sec) Normalized Throughput

TCP-SACK gives way to SCTP

16

SCTP TCP-SACK

slide-17
SLIDE 17

TCP-SACK gives way to SCTP

17

75 80 85 90 0.5 1 1.5 Simulation Time (sec) Normalized Throughput

SCTP TCP-SACK

slide-18
SLIDE 18

Conclusion

  • Irrespective of when a flow starts and bandwidth, SCTP
  • utperforms TCP-SACK and TCP-NewReno by

35%-41% (due to better loss recovery mechanisms)

  • TCP and SCTP traffic can grow together & TCP gives a

way to SCTP and vice versa even for the most imbalanced cases

18

Single-homed SCTP is TCP-friendly though it achieves higher throughput than TCP just as TCP- SACK or TCP-Reno perform better than TCP-Tahoe

slide-19
SLIDE 19

TCP-Friendliness of CMT*

* Experimental extension to SCTP (J. Iyengar, PEL @Univ. of Delaware, 2006)

slide-20
SLIDE 20

Motivation

20

  • TCP-F is defined for end-to-end transport connections
  • ver a single-path
  • J. Iyengar et. al. studied performance of CMT with the

assumption of bottleneck-independent topology

How does CMT behave when the tight link is shared between the CMT subflows and other TCP flows?

slide-21
SLIDE 21

Experimental Framework

21

100 Mbps 2 msec 100 Mbps 14 msec 100 Mbps 14 msec

  • n= 8, 16, 32, 48, 64 TCP flows from Si to Di in the background.

Then, introduce either (I), (II), or (III) (I) TCP1 (A1 to B1), TCP2 (A2 to B2) (II) SCTP1 (A1 to B1), SCTP2 (A2 to B2) (III) two-homed CMT where CMT-sub1 (A1 to B1), CMT-sub2 (A2 to B2)

  • RED queue @ tight link with minth= 5pks, maxth = 3*minth, wq =

0.002, maxp = 0.02, buffer size = BWxDelay

  • Metrics: Per-flow Throughput, Avg. flow Throughput, Fairness

Index

slide-22
SLIDE 22

22

Research Questions

  • What is the bandwidth share of two-homed

CMT compared to two independent TCP or SCTP flows?

  • What is the cost of introducing one two-

homed CMT flow into the network compared to two independent TCP or SCTP flows?

slide-23
SLIDE 23

23

Hypotheses

(I) Introducing two TCP flows: all TCP flows get an equal share of the bandwidth. (II) Introducing two SCTP flows: SCTP flows get >= share of the bandwidth compared to TCP flows. (III) Introducing one two-homed CMT flow*: CMT flow gets >= share of the bandwidth compared to two TCP or SCTP flows *CMT shares TSN space and ACK and more resilient to losses (J. Iyengar, 2006)

slide-24
SLIDE 24

24

Results

5 10 15 20 25 10 20 30 40 50 60 Throughput (Mbps) n (greedy TCP flows in the background) CMT SCTP1+SCTP2 TCP1+TCP2 2*bw/(n+2)

slide-25
SLIDE 25

25

Results

10 20 30 40 50 2 3 4 5 6 7 8 9 Throughput (Mbps) n (greedy TCP flows in the background) CMT SCTP1+SCTP2 2*bw/(n+2) 10 20 30 40 50 2 4 6 8 10 12 Throughput (Mbps) n (greedy TCP flows in the background) CMT SCTP1+SCTP2 2*bw/(n+2)

wq= 0.002 wq=0.001(RED reacts less aggressively

  • n bursty traffic)
slide-26
SLIDE 26

26

Results

2 4 6 8 10 10 20 30 40 50 60 Throughput (Mbps) n (greedy TCP flows in the background)

  • avg. w/ CMT
  • avg. w/ SCTP1+SCTP2
  • avg. w/ TCP1+TCP2

bw/(n+2)

slide-27
SLIDE 27

Conclusion

  • CMT <= two SCTP flows (due to bustier traffic

created by CMT)

  • CMT > two TCP flows (due to better loss recovery

and resilience to losses b/c of TSN space and ACK information)

  • CMT has AIMD-based congestion control which

allows other TCP flows to co-exist in the network

27

Two-homed CMT is TCP-friendly though it achieves higher throughput than two TCP flows just as two TCP-Reno flows would outperform two TCP-Tahoe flows

slide-28
SLIDE 28

Discussion and the End…

28

  • Other CMT-like schemes (CP, MulTFRC, mulTCP, MPAT,

PA-MulTCP, MPTCP, …)

  • Criticism to TCP-Friendliness (i.e, Flow-Rate Fairness) –

Cost Fairness (B. Briscoe)

  • TCP-F (or another fairness criteria) should include

multihoming and CMT!