Peer-to-Peer Networks 15 Self-Organization Christian Schindelhauer - - PowerPoint PPT Presentation

peer to peer networks
SMART_READER_LITE
LIVE PREVIEW

Peer-to-Peer Networks 15 Self-Organization Christian Schindelhauer - - PowerPoint PPT Presentation

Peer-to-Peer Networks 15 Self-Organization Christian Schindelhauer Technical Faculty Computer-Networks and Telematics University of Freiburg Topology-Management T-Man: Fast Gossip-based Construction of Large-Scale Overlay Topologies Mark


slide-1
SLIDE 1

Peer-to-Peer Networks

15 Self-Organization

Christian Schindelhauer

Technical Faculty Computer-Networks and Telematics University of Freiburg

slide-2
SLIDE 2

Topology-Management

T-Man: Fast Gossip-based Construction of Large-Scale Overlay Topologies Mark Jelasity Ozalp Babaoglu, 1994

2

slide-3
SLIDE 3

Distributed Topology Construction T-Man

3

slide-4
SLIDE 4

Finding a Torus

4

slide-5
SLIDE 5

Convergence of T-MAN

5

slide-6
SLIDE 6

T-Chord

Chord on demand, A Montresor, M Jelasity, O Babaoglu - Peer-to-Peer Computing, 2005 Apply self-organization to Chord

compare insertion operation Pastry

T-Chord

Apply T-Man preferring Chord edges

T-Chord-Prox

rank according to RTT

6

slide-7
SLIDE 7

1st rank nearest sucessor/predecessor on the 
 ring For each exponent select from view the nodes nearest to

Ranking Function T-Chord

7

[0, 2m − 1] j ∈ [1, m − 1] [ID + 2j mod 2m, ID + 2j+1 − 1 mod 2m]

slide-8
SLIDE 8

1st rank nearest sucessor/predecessor on the 
 ring For each exponent select from view the nodes nearest to measure latency (RTT) for p random nodes from view in such intervals and choose the closest

Ranking Function T-Chord-Prox

8

[0, 2m − 1] j ∈ [1, m − 1] [ID + 2j mod 2m, ID + 2j+1 − 1 mod 2m]

slide-9
SLIDE 9

Adaption for Chord

9

slide-10
SLIDE 10

After Exchange of Links

10

slide-11
SLIDE 11

T-Chord Performance

Starting with a neighbors on the ring Loss rate and hop count experiments on a real-word dataset from 2002

11

1e-04 0.001 0.01 0.1 1 10 100 5 10 15 20 2 3 4 5 6 7 8 Loss Rate (%) Hop Count (n.) Cycles T-Chord (%) T-Chord-Prox (%) T-Chord (n.) T-Chord-Prox (n.)

slide-12
SLIDE 12

T-Chord Performance

Starting with a neighbors on the ring Loss rate and hop count

12

4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 210 211 212 213 214 215 216 217 218 Hop Count Size Chord T-Chord T-Chord-Prox

slide-13
SLIDE 13

T-Chord Performance

Message Delay

13

200 300 400 500 600 700 800 210 211 212 213 214 215 216 217 218 Latency (ms) Size Chord T-Chord T-Chord-Prox

slide-14
SLIDE 14

T-Chord Performance

Robustness

14

500 1000 1500 2000 2500 3000 3500 10 20 30 40 50 Message Delay (ms) Crashed nodes (%) Chord (crash) T-Chord (crash) T-Chord (churn) T-Chord-Prox (churn)

2 4 6 8 10 12 14 10 20 30 40 50 Hop Count Crashed nodes (%) Chord (crash) T-Chord (crash) T-Chord (churn) T-Chord-Prox (churn) 1 2 3 4 5 6 7 8 10 20 30 40 50 Loss rate (%) Crashed nodes (%) Chord (crash) T-Chord (crash) T-Chord (churn) T-Chord-Prox (churn)

slide-15
SLIDE 15

Peer-to-Peer Networks

15 Self-Organization

Christian Schindelhauer

Technical Faculty Computer-Networks and Telematics University of Freiburg