BitTorrent Experiments on Testbeds: A Study of the Impact ... · BitTorrent Experiments on...
Transcript of BitTorrent Experiments on Testbeds: A Study of the Impact ... · BitTorrent Experiments on...
Title Introduction Methodology Experimental Results Conclusion Backup
BitTorrent Experiments on Testbeds: A
Study of the Impact of Network Latencies
Ashwin Rao, Arnaud Legout, and Walid Dabbous
INRIA, Projet Planete(ashwin.rao,arnaud.legout,walid.dabbous)@inria.fr
1/17
Title Introduction Methodology Experimental Results Conclusion Backup
Outline
Introduction
Methodology
Experimental ResultsHomogeneous LatencyHeterogeneous Latency
Conclusion
2/17
Title Introduction Methodology Experimental Results Conclusion Backup
Overview of BitTorrent
3/17
Title Introduction Methodology Experimental Results Conclusion Backup
Overview of BitTorrent
3/17
Title Introduction Methodology Experimental Results Conclusion Backup
Overview of BitTorrent
3/17
Title Introduction Methodology Experimental Results Conclusion Backup
Overview of BitTorrent
3/17
Title Introduction Methodology Experimental Results Conclusion Backup
Overview of BitTorrent
3/17
Title Introduction Methodology Experimental Results Conclusion Backup
Overview of BitTorrent
3/17
Title Introduction Methodology Experimental Results Conclusion Backup
Evaluation of BitTorrent Performance
Planetlab sites [planete-lab.org]Grid5000 sites[grid5000.fr]
• Grid5000 and PlaneteLab Testbeds
• Absence of network latency between
• Instances of application running on same machine• Machines in the same LAN
Does network latency affect the outcome of BitTorrentexperiments performed on testbeds
4/17
Title Introduction Methodology Experimental Results Conclusion Backup
Evaluation of BitTorrent Performance
Planetlab sites [planete-lab.org]Grid5000 sites[grid5000.fr]
• Grid5000 and PlaneteLab Testbeds• Absence of network latency between
• Instances of application running on same machine• Machines in the same LAN
Does network latency affect the outcome of BitTorrentexperiments performed on testbeds
4/17
Title Introduction Methodology Experimental Results Conclusion Backup
Evaluation of BitTorrent Performance
Planetlab sites [planete-lab.org]Grid5000 sites[grid5000.fr]
• Grid5000 and PlaneteLab Testbeds• Absence of network latency between
• Instances of application running on same machine• Machines in the same LAN
Does network latency affect the outcome of BitTorrentexperiments performed on testbeds
4/17
Title Introduction Methodology Experimental Results Conclusion Backup
Can Latency Impact BitTorrent Performance?
TCP Ramp-up and Impact ofUpload Rates
Delays in receivingBitTorrent control
messages
Can testbeds such as Grid5000 be used for experimentalevaluation of BitTorrent?
5/17
Title Introduction Methodology Experimental Results Conclusion Backup
Can Latency Impact BitTorrent Performance?
TCP Ramp-up and Impact ofUpload Rates
Delays in receivingBitTorrent control
messages
Can testbeds such as Grid5000 be used for experimentalevaluation of BitTorrent?
5/17
Title Introduction Methodology Experimental Results Conclusion Backup
Can Latency Impact BitTorrent Performance?
TCP Ramp-up and Impact ofUpload Rates
Delays in receivingBitTorrent control
messages
Can testbeds such as Grid5000 be used for experimentalevaluation of BitTorrent?
5/17
Title Introduction Methodology Experimental Results Conclusion Backup
Outline
Introduction
Methodology
Experimental ResultsHomogeneous LatencyHeterogeneous Latency
Conclusion
6/17
Title Introduction Methodology Experimental Results Conclusion Backup
Testbed Setup
4 Machines of Grid5000 Experimental Testbed
7/17
Title Introduction Methodology Experimental Results Conclusion Backup
Testbed Setup
Machines capable of running 100 instances of BitTorrent Client
7/17
Title Introduction Methodology Experimental Results Conclusion Backup
Testbed Setup
Peers on same machine communicate using Loopback device
7/17
Title Introduction Methodology Experimental Results Conclusion Backup
Testbed Setup
Peers on different machines communicate using Ethernetdevice
7/17
Title Introduction Methodology Experimental Results Conclusion Backup
Testbed Setup
Tracker and Seed placed on same machine
7/17
Title Introduction Methodology Experimental Results Conclusion Backup
Experiment Setup
• Experiment Parameters• 50 MB file• 1 Tracker, 1 Seed, and 300 Leechers• Upload rates - 10 KiB/s to 100 KiB/s• Emulate RTT - 1000 ms
• Metric• Download completion time
8/17
Title Introduction Methodology Experimental Results Conclusion Backup
Outline
Introduction
Methodology
Experimental ResultsHomogeneous LatencyHeterogeneous Latency
Conclusion
9/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Setup
Same delay on Loopback and Ethernet Device
10/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Experiment Results
Download Completion Time vs RTT
11/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Experiment Results
1000 ms RTT does not increase Average Download CompletionTime by more than 15%
11/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Experiment Results
Download Completion Time not a monotonously increasingfunction of RTT
11/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Experiment Results
Download Completion Time not a monotonously increasingfunction of RTT
11/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Experiment Results
Impact of latency not observed when seed is fast
11/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Experiment Results
Marginal impact of RTT
11/17
Title Introduction Methodology Experimental Results Conclusion Backup
Homogeneous Latency - Experiment Results
Marginal impact of RTT
11/17
Title Introduction Methodology Experimental Results Conclusion Backup
Heterogeneous Latency
Each machine used to abstract an AS
12/17
Title Introduction Methodology Experimental Results Conclusion Backup
Heterogeneous Latency
Different latency on loopback and ethernet device
12/17
Title Introduction Methodology Experimental Results Conclusion Backup
Results - Small RTT Between Peers
Upload Rate limited to20 kB/s.
RTT between a pair ofleechers.
RTT between any two peers is less than 100 ms
13/17
Title Introduction Methodology Experimental Results Conclusion Backup
Results - Small RTT Between Peers
Upload Rate limited to20 kB/s.
RTT between a pair ofleechers.
RTT between any two peers is less than 100 ms
Marginal Impact of Latency
13/17
Title Introduction Methodology Experimental Results Conclusion Backup
Results - Small RTT Between Peers
Upload Rate limited to20 kB/s.
RTT between a pair ofleechers.
RTT between any two peers is less than 100 ms
Marginal Impact of Latency
13/17
Title Introduction Methodology Experimental Results Conclusion Backup
Results - Large RTT Between Some Peers
Upload Rate limited to20 kB/s.
RTT between a pair ofleechers.
RTT between some of the peers is greater than 200 ms
14/17
Title Introduction Methodology Experimental Results Conclusion Backup
Results - Large RTT Between Some Peers
Upload Rate limited to20 kB/s.
RTT between a pair ofleechers.
RTT between some of the peers is greater than 200 ms
Marginal Impact of Latency
14/17
Title Introduction Methodology Experimental Results Conclusion Backup
Outline
Introduction
Methodology
Experimental ResultsHomogeneous LatencyHeterogeneous Latency
Conclusion
15/17
Title Introduction Methodology Experimental Results Conclusion Backup
Conclusion
• Latency has a marginal impact on the outcome ofBitTorrent experiments
BitTorrent experiments can be performed on testbedswithout explicitly emulating latency.
16/17
Title Introduction Methodology Experimental Results Conclusion Backup
Conclusion
• Latency has a marginal impact on the outcome ofBitTorrent experiments
BitTorrent experiments can be performed on testbedswithout explicitly emulating latency.
16/17
Title Introduction Methodology Experimental Results Conclusion Backup
BitTorrent Experiments on Testbeds: A Study of
the Impact of Network Latencies
Ashwin Rao ([email protected])
17/17
Title Introduction Methodology Experimental Results Conclusion Backup
TCP Segmentation Offloading
0 0.2 0.4 0.6 0.8
1
1 10 100 1000 10000
TCP Payload Length (bytes)
TSO Disabled
0 ms100 ms400 ms
0 0.2 0.4 0.6 0.8
1
1 10 100 1000 10000
CD
F
WIDE backbone
WIDE backbone
0 0.2 0.4 0.6 0.8
1
1 10 100 1000 10000
TSO Enabled
0 ms100 ms400 ms
Figure: Impact of Segmentation Offloading
−/16
Title Introduction Methodology Experimental Results Conclusion Backup
Number of Nodes on a Machine
0
100
200
300
400
500
600
700
800
900
1000
1100
0 25 50 75 100
RT
T e
stim
ate
of
TC
P (
ms)
Number of leechers running on a given machine
Minimum RTT 0 msMinimum RTT 400 ms
Minimum RTT 1000 ms
Figure: Impact of Number of Nodes on a Machine
Machines support up to 100 instances of a leecher−/16
Title Introduction Methodology Experimental Results Conclusion Backup
Send Call Inter-arrival Time
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
<20 40 70 90 120 190 275 420 550 >1000
CD
F
Inter arrival time of data messages on a given connection (ms)
10 kB/s20 kB/s50 kB/s
100 kB/s
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
<20 40 70 90 120 190 275 420 550 >1000
CD
FInter arrival time of data messages
on a given connection (ms)
10 kB/s20 kB/s50 kB/s
100 kB/s
−/16
Title Introduction Methodology Experimental Results Conclusion Backup
RTT between pair of peersAS1 AS2 AS3 AS4 AS5
AS1 8 ms 40 ms 60 ms 210 ms 210 msAS2 40 ms 20 ms 80 ms 230 ms 230 msAS3 60 ms 80 ms 40 ms 250 ms 250 msAS4 210 ms 230 ms 250 ms 100 ms 400 msAS5 210 ms 230 ms 250 ms 400 ms 200 ms
Table: RTT between a pair of leechers.
AS1 AS2 AS3 AS4 AS5
AS ′1 20 ms 40 ms 60 ms 210 ms 210 ms
AS ′2 40 ms 60 ms 80 ms 230 ms 230 ms
AS ′3 60 ms 80 ms 100 ms 250 ms 250 ms
AS ′4 210 ms 230 ms 250 ms 400 ms 400 ms
AS ′5 210 ms 230 ms 250 ms 400 ms 400 ms
Table: RTT between the initial seed and the leechers. −/16
Title Introduction Methodology Experimental Results Conclusion Backup
Results - Large RTT Between Some Peers
Figure: Upload rate limited to 50 KiB/s. RTT between some peersmore than 100 ms.
RTT of even400 ms hasmarginal impact ondownloadcompletion time
AS1 AS3 AS5
AS1 8 ms 60 ms 210 msAS3 60 ms 40 ms 250 msAS5 210 ms 250 ms 200 msAS ′
1 20 ms 60 ms 210 msAS ′
5 210 ms 250 ms 400 ms
Table: RTT between a pair of leechers. −/16
Title Introduction Methodology Experimental Results Conclusion Backup
Results - Large RTT Between Some Peers
Figure: Upload rate limited to 50 KiB/s. RTT between some peersmore than 100 ms.
RTT of even400 ms hasmarginal impact ondownloadcompletion time
AS1 AS3 AS5
AS1 8 ms 60 ms 210 msAS3 60 ms 40 ms 250 msAS5 210 ms 250 ms 200 msAS ′
1 20 ms 60 ms 210 msAS ′
5 210 ms 250 ms 400 ms
Table: RTT between a pair of leechers. −/16