Click to edit Master title style Click to edit Master title style - - PowerPoint PPT Presentation

▶
click to edit master title style click to edit master
SMART_READER_LITE
LIVE PREVIEW

Click to edit Master title style Click to edit Master title style - - PowerPoint PPT Presentation

Oct 16, 2023 369 likes •569 views

Click to edit Master title style Click to edit Master title style Click to edit Master title style ? Click to edit Master text styles ? ? Click to edit Master text styles ? Click to edit Master text styles TAROC: TCP-Aware RObust ` Second

slide-1
SLIDE 1

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

1

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

1

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

1

?

`

?

` ?

1 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

TAROC: TCP-Aware RObust Header Compression Scheme

http://www.dmn.tzi.org/ietf/rohc/draft-ietf-rohc-tcp-taroc-01.txt

HongBin Liao, Qian Zhang, Wenwu Zhu, Ya-Qin Zhang Microsoft Research

slide-2
SLIDE 2

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

2

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

2

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

2

?

`

?

` ?

2 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Outline

Basic concepts of TAROC

robustness of TAROC (window-based LSB encoding) efficiency of TAROC (TCP congestion window tracking)

Improvements in this version

more accurate for TCP congestion window tracking more efficient for bulk data transmission adjust packet header format

Tracking / compression performance analysis

slide-3
SLIDE 3

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

3

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

3

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

3

?

`

?

` ?

3 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

0-100 100-200 200-300 300-400 400-500 800-900 600-700 1400-1500 500-600 900-1000 1000-1100 1100-1200 1200-1300 1300-1400 700-800

  • thers

Seqno variation distribution

0-100 100-200 200-300 300-400 400-500 500-600 600-700 700-800 800-900 900-1000 1000-1100 1100-1200 1200-1300 1300-1400 1400-1500 1500-1600 1600-1700 1700-1800 1800-1900 1900-2000 2000-2100 2100-2200 2200-2300 2300-2400 2400-2500 2500-2600 2600-2700 2700-2800 2899-2900 2900-3000 3000-3100 3100-3200 3200-3300 3300-3400 3400-3500 3500-3600 3600-3700 3700-3800 3800-3900 3900-4000 4000-4100 4100-4200 4200-4300 4300-4400 4400-4500 4500-4600 4600-4700 4700-4800 4800-4900 4900-5000

  • thers

Window-based LSB encoding

Ackno variation distribution 03/11/2001 - 03/17/2001 SDSC : San Diego Supercomputer Center NLANR network traffic packet header traces

slide-4
SLIDE 4

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

4

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

4

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

4

?

`

?

` ?

4 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Tracking-based TCP congestion window estimation

Reconstruct the congestion control behavior of TCP sender at the compressor Adjust congestion window size in each state based on seqno /ackno

slide-5
SLIDE 5

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

5

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

5

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

5

?

`

?

` ?

5 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Several issues of TAROC version 0

Accuracy of tracking-based TCP congestion window estimation

Ackno-only window tracking Receiver advertised window driven case Multiple link case

Efficiency of window-based LSB coding

Bulk file transmission usually has fixed payload size Irregular seqno / ackno for other traffics

More realistic and abundant simulation conditions

Simulation topologies (multiple bottleneck, multiple link) Simulation BER (10-5 ~ 10-8)

slide-6
SLIDE 6

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

6

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

6

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

6

?

`

?

` ?

6 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Seqno / Ackno-based congestion window estimation RTT-smoothed rate estimation

new packet comes in

Estimated congestion window adjustment

Congestion window increase seqno/ackno-based congestion window estimation Congestion window shrunk seqno/ackno-based congestion window estimation RTT-smoothed rate estimation (TCP linear / sub-linear increasing) Congestion window increase seqno/ackno-based congestion window estimation Congestion window shrunk seqno/ackno-based congestion window estimation RTT-smoothed rate estimation (TCP linear / sub-linear increasing) new packet comes in

Improvement in congestion window estimation

slide-7
SLIDE 7

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

7

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

7

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

7

?

`

?

` ?

7 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

RTT estimation

Packet clustering distribution

RTT-smoothed rate estimation

  • ×

> µ

  • i

i

  • ∑

=

Congestion window shrunk

Check the conflict between rate variation and cwnd variation

RTT-smoothed rate calculation

`sub-clustering `near-uniform distribution

Inter-RTT variation judgment

slide-8
SLIDE 8

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

8

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

8

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

8

?

`

?

` ?

8 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Compressed TCP/IP header structure

  • F
  • 1

2 3

  • Almost the same as the one

in version 1 except for

Bulk data transmission (fixed payload size) <F+WAS fields>

W-LSB encoding applied to

Window, Acknowledgement Number, Sequence Number and IP-ID TS value, TS echo reply in TCP Timestamp option

Delta coding applied to

SACK option

slide-9
SLIDE 9

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

9

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

9

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

9

?

`

?

` ?

9 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

WAS field

PADDING C_IP_ID C_SEQ C_ACK C_WIN PT

3~7 6 14 14 13 101 2~5 3 11 11 10 00x 111 110 100 011 010 xxx PT value 4~8 13 14 14 13 4~9 10 18 18 17 5~11 14 22 22 21 3~5 8 9 9 8 2~4 5 6 6 5 1~2 3 4 4 3 1 length of WASI (bytes) C_IP_ID length (bits) C_SEQ length (bits) C_ACK length (bits) C_WIN length (bits) F

slide-10
SLIDE 10

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

10

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

10

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

10

?

`

?

` ?

10 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Simulation topologies

Single link Multiple links

Performance analysis

sender base station receiver

8Mbps, 100ms queue size = 6 bw, 100ms

sender base stations receiver

8 M b p s , 1 m s queue size = 6 bw, 100ms

1 5 Simulation conditions

Tested header compression schemes

` no compression (None) ` RFC 1144 (VJHC) (4 bytes) ` RFC 2507 (IPHC) (5 bytes) ` our scheme (TAROC) (5 bytes) ` no error happens with header size (4 bytes) (Ideal)

BER in wireless channel (10-9 to 10-5) Bandwidth in wireless channel (9.6Kbps to 384Kbps) TCP versions: Tahoe, Reno, Sack Buffer management: Droptail, RED, RED-ECN

slide-11
SLIDE 11

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

11

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

11

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

11

?

`

?

` ?

11 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Performance analysis

Effectiveness of rate tracking

` single bottleneck ` multiple bottlenecks Senders (8) Receivers (8)

1.5Mbps, 45ms queue size = 25 10Mbps, 2ms 1 M b p s , 3 m s

¡ - ¡ -

10Mbps, 5ms 1.5Mbps, 50ms 10Mbps, 5ms 10Mbps, 5ms

S0 S1 S2 D1 S3 D2 D3 D0

slide-12
SLIDE 12

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

12

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

12

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

12

?

`

?

` ?

12 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Accuracy of congestion window tracking

Tracking cwnd using seqno

2 4 6 8

slide-13
SLIDE 13

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

13

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

13

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

13

?

`

?

` ?

13 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Accuracy of congestion window tracking

Tracking cwnd using ackno

2 4 6 8

slide-14
SLIDE 14

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

16

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

16

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

16

?

`

?

` ?

16 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

1143 1134 1118 1045 1055 996 Throughput (Byte/s) 296 3*10-6 14.76 13.86 12.25 4.92 5.92 Improvement Percentage (%)

  • 0.69

852 9.96 1126 12.20 1159 12.75 1167 12.84 1169 12.93 1170

IPHC

14.78 14.30 9.50

  • 0.88

Improvement Percentage (%) 1056 1047 1003 835 916 Throughput (Byte/s) 296 10-5 13.87 13.48 12.60 11.13 Improvement Percentage (%) 1166 1162 1153 1138 1024 Throughput (Byte/s) 296 10-6 13.75 13.36 12.88 12.49 Improvement Percentage (%) 1175 1171 1166 1162 1033 Throughput (Byte/s) 296 3*10-7 13.82 13.43 12.95 13.33 Improvement Percentage (%) 1178 1174 1169 1173 1035 Throughput (Byte/s) 296 10-7 13.80 13.42 12.93 13.42 Improvement Percentage (%) 1179 1175 1170 1175 1036 Throughput (Byte/s) 296 3*10-8 13.8 13.42 13.03 13.51 Improvement Percentage (%) 1179 1175 1171 1176 1036 Throughput (Byte/s) 296 10-8

Ideal TAROC TWICE VJHC None Performance MTU (Byte) BER

Performance comparison: 9.6kb/s for Reno

slide-15
SLIDE 15

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

17

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

17

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

17

?

`

?

` ?

17 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

  • 1
  • Performance comparison: 9.6kb/s for Reno
slide-16
SLIDE 16

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

18

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

18

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

18

?

`

?

` ?

18 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

9194 9136 8510 5913 6528 8721 Throughput (Byte/s) 576 3*10-6 5.42 4.76

  • 3.42
  • 32.20
  • 25.15

Improvement Percentage (%)

  • 33.09

3045

  • 17.25

9189 4.19 11426 1.77 12373 3.86 12720 4.56 12834

IPHC

5.34 5.38

  • 3.05
  • 34.63

Improvement Percentage (%) 4794 4796 4412 2975 4551 Throughput (Byte/s) 576 10-5 5.24 5.20 1.63

  • 10.49

Improvement Percentage (%) 11687 11682 11286 9940 11105 Throughput (Byte/s) 576 10-6 4.98 4.79 3.56

  • 1.12

Improvement Percentage (%) 12520 12497 12351 11792 11926 Throughput (Byte/s) 576 3*10-7 5.1 4.83 4.54 3.03 Improvement Percentage (%) 12778 12745 12710 12526 12158 Throughput (Byte/s) 576 10-7 5.0 4.72 4.74 4.46 Improvement Percentage (%) 12859 12825 12828 12793 12247 Throughput (Byte/s) 576 3*10-8 4.95 4.71 4.78 4.76 Improvement Percentage (%) 12882 12852 12861 12859 12274 Throughput (Byte/s) 576 10-8

Ideal TAROC TWICE VJHC None Performance MTU (Byte) BER

Performance comparison: 114kb/s for Reno

slide-17
SLIDE 17

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

19

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

19

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

19

?

`

?

` ?

19 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

Performance comparison: 114kb/s for Reno

2 3 4 5 6 7 8 9

slide-18
SLIDE 18

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

20

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

20

Click to edit Master title style

? Click to edit Master text styles

` Second level

? Third level

` Fourth level ? Fifth level

20

?

`

?

` ?

20 draft-ietf-rohc-tcp-taroc-01.txt Mar 2001, 50th IETF

11425 11445 10409 7035 7714 10785 Throughput (Byte/s) 576 3*10-6 5.93 6.12

  • 3.48
  • 34.8
  • 28.5

Improvement Percentage (%)

  • 32.7

3361

  • 38.3

12418

  • 21.0

18547

  • 5.22

22090 1.95 23574 4.02 24007

IPHC

5.72 5.72

  • 3.18
  • 33.4

Improvement Percentage (%) 5282 5282 4837 3326 4996 Throughput (Byte/s) 576 10-5 3.57 2.17

  • 3.92
  • 27.0

Improvement Percentage (%) 20843 20561 19337 14682 20125 Throughput (Byte/s) 576 10-6 3.32 3.1

  • 0.06
  • 12.8

Improvement Percentage (%) 24264 24211 23468 20472 23483 Throughput (Byte/s) 576 3*10-7 4.63 4.49 3.52

  • 1.36

Improvement Percentage (%) 24388 24355 24130 22990 23308 Throughput (Byte/s) 576 10-7 5.17 5.11 4.83 3.36 Improvement Percentage (%) 24320 24305 24242 23901 23124 Throughput (Byte/s) 576 3*10-8 5.24 5.11 5.03 4.48 Improvement Percentage (%) 24288 24257 24238 24113 23078 Throughput (Byte/s) 576 10-8

Ideal TAROC TWICE VJHC None Performance MTU (Byte) BER

Performance comparison: 384kb/s for Reno