Can Realistic BitTorrent Experiments Be Performed on Clusters? - - PowerPoint PPT Presentation

▶
can realistic bittorrent experiments be performed on
SMART_READER_LITE
LIVE PREVIEW

Can Realistic BitTorrent Experiments Be Performed on Clusters? - - PowerPoint PPT Presentation

Aug 05, 2023 157 likes •859 views

Can Realistic BitTorrent Experiments Be Performed on Clusters? Ashwin Rao , Arnaud Legout, and Walid Dabbous INRIA, Project Plan` ete (ashwin.rao,arnaud.legout,walid.dabbous)@inria.fr 1 Overview of BitTorrent 2 Overview of BitTorrent 2

slide-1
SLIDE 1

Can Realistic BitTorrent Experiments Be Performed on Clusters?

Ashwin Rao, Arnaud Legout, and Walid Dabbous

INRIA, Project Plan` ete (ashwin.rao,arnaud.legout,walid.dabbous)@inria.fr

1

slide-2
SLIDE 2

Overview of BitTorrent

2

slide-3
SLIDE 3

Overview of BitTorrent

2

slide-4
SLIDE 4

Overview of BitTorrent

2

slide-5
SLIDE 5

Overview of BitTorrent

2

slide-6
SLIDE 6

Overview of BitTorrent

2

slide-7
SLIDE 7

Overview of BitTorrent

2

slide-8
SLIDE 8

Overview of BitTorrent

2

slide-9
SLIDE 9

Platforms for BitTorrent Experiments

PlanetLab sites [planet-lab.org]

3

slide-10
SLIDE 10

Platforms for BitTorrent Experiments

PlanetLab sites [planet-lab.org] Clusters

3

slide-11
SLIDE 11

Platforms for BitTorrent Experiments

PlanetLab sites [planet-lab.org] Clusters PlanetLab Clusters Reproducibility X

  • Scalability

X

  • 3
slide-12
SLIDE 12

Platforms for BitTorrent Experiments

PlanetLab sites [planet-lab.org] Clusters PlanetLab Clusters Reproducibility X

  • Scalability

X

  • RTT & Packet Loss
  • ?

3

slide-13
SLIDE 13

Platforms for BitTorrent Experiments

PlanetLab sites [planet-lab.org] Clusters PlanetLab Clusters Reproducibility X

  • Scalability

X

  • RTT & Packet Loss
  • ?

Experiments on Clusters Without Emulating RTT & Packet Loss?

3

slide-14
SLIDE 14

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics

4

slide-15
SLIDE 15

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics

4

slide-16
SLIDE 16

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics

4

slide-17
SLIDE 17

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics

4

slide-18
SLIDE 18

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics
  • Pipeline Requests
  • Time between decisions ≈ 10 seconds

4

slide-19
SLIDE 19

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics
  • Pipeline Requests
  • Time between decisions ≈ 10 seconds
  • Throughput

4

slide-20
SLIDE 20

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics
  • Pipeline Requests
  • Time between decisions ≈ 10 seconds
  • Throughput

Throughput ∝ 1 RTT √ Loss Rate

4

slide-21
SLIDE 21

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics
  • Pipeline Requests
  • Time between decisions ≈ 10 seconds
  • Throughput
  • Users limit Upload Rate

4

slide-22
SLIDE 22

Impact of RTT and Packet Loss

  • BitTorrent Protocol Dynamics
  • Pipeline Requests
  • Time between decisions ≈ 10 seconds
  • Throughput
  • Users limit Upload Rate

Need to Validate Impact

  • f RTT & Packet Loss

4

slide-23
SLIDE 23

Outline

Motivation

  • Experiment Setup
  • Impact of RTT
  • Impact of Packet Loss

5

slide-24
SLIDE 24

Testbed Setup

4 Machines of a Cluster

6

slide-25
SLIDE 25

Testbed Setup

A machine can have 100 instances of Real BitTorrent Client 1 tracker, 1 seed, 300 leechers

6

slide-26
SLIDE 26

Testbed Setup

Loopback device: Peers on the same machine

6

slide-27
SLIDE 27

Testbed Setup

Ethernet device: Peers on adjacent machines

6

slide-28
SLIDE 28

Testbed Setup

Emulate Same RTT and Packet Loss Rate

6

slide-29
SLIDE 29

Experiment Parameters and Metric

iPlane Measurement: 95% Paths

  • RTT ≤ 400 ms
  • Packet Loss Rate ≤ 5%

iPlane: An Information Plane for Distributed Services.

  • H. Madhyastha et al., OSDI 2006.

http://iplane.cs.washington.edu/

7

slide-30
SLIDE 30

Experiment Parameters and Metric

iPlane Measurement: 95% Paths

  • RTT ≤ 400 ms
  • Packet Loss Rate ≤ 5%

Experiment Parameters

  • RTT - up to 1000 ms
  • Packet Loss Rate - up to 5%
  • 50 MB file

Metric

  • Download completion time

7

slide-31
SLIDE 31

Outline

Motivation Experiment Setup

  • Impact of RTT
  • Impact of Packet Loss

8

slide-32
SLIDE 32

Impact of RTT - Experiment Results

9

slide-33
SLIDE 33

Impact of RTT - Experiment Results

9

slide-34
SLIDE 34

Impact of RTT - Experiment Results

9

slide-35
SLIDE 35

Impact of RTT - Experiment Results

9

slide-36
SLIDE 36

Impact of RTT - Experiment Results

9

slide-37
SLIDE 37

Impact of RTT - Experiment Results

9

slide-38
SLIDE 38

Impact of RTT - Experiment Results

9

slide-39
SLIDE 39

Impact of RTT - Experiment Results

Scenario Fast Seed and Slow Leechers

9

slide-40
SLIDE 40

Impact of RTT - Experiment Results

Scenario of Fast Peers

9

slide-41
SLIDE 41

Impact of RTT - Experiment Results

9

slide-42
SLIDE 42

Impact of RTT - Experiment Results Marginal Impact of RTT

9

slide-43
SLIDE 43

Outline

Motivation Experiment Setup Impact of RTT

  • Impact of Packet Loss

10

slide-44
SLIDE 44

Impact of Packet Loss (20 kB/s)

11

slide-45
SLIDE 45

Impact of Packet Loss (20 kB/s)

11

slide-46
SLIDE 46

Impact of Packet Loss (20 kB/s)

11

slide-47
SLIDE 47

Impact of Packet Loss (20 kB/s)

11

slide-48
SLIDE 48

Impact of Packet Loss (20 kB/s)

11

slide-49
SLIDE 49

Impact of Packet Loss (20 kB/s)

11

slide-50
SLIDE 50

Impact of Packet Loss (20 kB/s) Marginal Impact of Packet Loss

11

slide-51
SLIDE 51

Impact of Packet Loss (100 kB/s)

12

slide-52
SLIDE 52

Impact of Packet Loss (100 kB/s)

12

slide-53
SLIDE 53

Impact of Packet Loss (100 kB/s)

12

slide-54
SLIDE 54

Impact of Packet Loss (100 kB/s) Marginal Impact of Packet Loss

12

slide-55
SLIDE 55

Conclusions

  • Marginal Impact of RTT
  • Marginal Impact of Packet Loss
  • Avoid Emulation of RTT and Packet Loss

13

slide-56
SLIDE 56

Conclusions

  • Marginal Impact of RTT
  • Marginal Impact of Packet Loss
  • Avoid Emulation of RTT and Packet Loss

Dedicated clusters can be used for BitTorrent Experiments

13

slide-57
SLIDE 57

Can Realistic BitTorrent Experiments Be Performed on Clusters?

Questions ashwin.rao@inria.fr

14

slide-58
SLIDE 58

Can Realistic BitTorrent Experiments Be Performed on Clusters?

Thank you! ashwin.rao@inria.fr

14

slide-59
SLIDE 59

BACKUPS

Variable Latency .torrent file TCP SQRT B1

slide-60
SLIDE 60

Heterogeneous Latency - Setup

Different latency added on Loopback and Ethernet Devices

BACKUP B2

slide-61
SLIDE 61

Heterogeneous Latency - Setup

Grid’5000 Machine to Abstract an AS

BACKUP B2

slide-62
SLIDE 62

Results - Small RTT Between Peers

Upload Rate limited to 20 kB/s. RTT between a pair of leechers. RTT between any two peers is less than 100 ms

BACKUP B3

slide-63
SLIDE 63

Results - Small RTT Between Peers

Upload Rate limited to 20 kB/s. RTT between a pair of leechers. RTT between any two peers is less than 100 ms

BACKUP B3

slide-64
SLIDE 64

Results - Small RTT Between Peers

Upload Rate limited to 20 kB/s. RTT between a pair of leechers. RTT between any two peers is less than 100 ms

BACKUP B3

slide-65
SLIDE 65

Results - Large RTT Between Some Peers

Upload Rate limited to 20 kB/s. RTT between a pair of leechers. RTT between some of the peers is greater than 200 ms

BACKUP B4

slide-66
SLIDE 66

Results - Large RTT Between Some Peers

Upload Rate limited to 20 kB/s. RTT between a pair of leechers. RTT between some of the peers is greater than 200 ms

BACKUP B4

slide-67
SLIDE 67

.torrent File

info dictionary describing files in torrent name File Name length File size in bytes piece length Size of a piece pieces 20-byte hash values; one per piece announce The announce URL of tracker ...

BACKUP B5

slide-68
SLIDE 68

TCP Square Root

Throughput = MSS RTT C √p where MSS Maximum Segment Size RTT Round trip time p Loss rate C constant depending on TCP implementation

BACKUP B6