breaking the single-path barrier

21
Breaking the Single- Path Barrier Brad Smith Jack Baskin SoE Research Review Day 10/20/2011

Upload: evania

Post on 23-Feb-2016

57 views

Category:

Documents


0 download

DESCRIPTION

Breaking the Single-Path Barrier. Brad Smith Jack Baskin SoE Research Review Day 10/20/2011. Outline. Research Corporate Partnership Open Source Network Lab. The Internet is Single-Path. Compute best path per destination Destination-based (“hop-by-hop”) forwarding 2 problems - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Breaking the Single-Path Barrier

Breaking the Single-Path Barrier

Brad SmithJack Baskin SoE Research Review Day

10/20/2011

Page 2: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

2

Outline

• Research• Corporate Partnership• Open Source Network Lab

10/20/11

Page 3: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

3

The Internet is Single-Path

• Compute best path per destination• Destination-based (“hop-by-hop”) forwarding• 2 problems– Quality-of-Service– Congestion

S

10/20/11

Page 4: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

4

Problem – Quality-of-Service

• Requires multiple paths per destination• Example – 2 paths

• 2Mbps, 200ms• 100Kbps, 20ms

• Depends on application!– Video streaming– Voice over IP (VoIP)

b/w = 100Kbpslatency = 20ms

b/w = 2Mbpslatency = 200ms

Next hop for D?S

a

b

D

10/20/11

Page 5: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

5

Problem – Congestion

45

45

45

45

1616 32

32

16

16

18 18

321616

18

18

16

16

3232

32

32

32

7

7

74

4 7

4

7

4

7

7

7

10/20/11

Strong tendency for paths to share links…

Page 6: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

6

Previous Work

• Circuit switch – special path per flow• Solve part of the problem– Congestion – only need a small number (≤ 4)– Minimum delay– Partial solutions - disjoint widest and shortest

• In practice - over-provision a single path• Challenge – what are enough paths?

10/20/11

Page 7: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

7

Best Set of Paths• Paths as points in multi-dimensional space• Some paths are “better” than others• Best set of paths are those with none better• Paths that provide the full range of performance

100Kbps, 20ms100Kbps, 20ms 100Kbps, 20ms

10/20/11

Page 8: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

8

Using the Best Set of Paths

• Assign flows to paths that satisfy QoS• In general, there is more than one…• …choose one that minimizes congestion

10/20/11

100Kbps, 20ms

Page 9: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

9

Simulations

• Generate random networks• Compute routing tables at all nodes– Total bandwidth– Delay

• Generate random stream of flows• Use oracle to assign flows to paths– Satisfies QoS– Has bandwidth

10/20/11

Page 10: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

10

Simulations (cont)

• Measure Call Acceptance Ratio– % flow requests successfully routed

• Across range of networks– Size – number of vertices– Connectivity – average degree (# neighbors)

10/20/11

Page 11: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

11

CAR: 350 Vertices, Degree 32

10/20/11

Page 12: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

12

CAR: 350 Vertices, Degree 16

10/20/11

Page 13: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

13

CAR: 350 Vertices, Degree 4

10/20/11

Page 14: Breaking the Single-Path Barrier

14

• Converge when lightly loaded and over-loaded

• Multipath does better in-between

• Fewer resources poorer and less distinct performance

10/20/11 Jack Baskin SoE Research Review Day

Page 15: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

15

Target CAR Flow Rate

10/20/11

95%

600Load supported by given infrastructure & routing architecture.

Page 16: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

16

95% CAR Rate: Degree 32

10/20/11

Multipath provides dramatic capacity increase with same infrastructure.

Page 17: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

17

MP @ 95% CAR Rate – # Vertices

10/20/11

Solid gains with increasing infrastructure.

Page 18: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

18

Ratio MP:SP 95% CAR – # Vertices

10/20/11

4 to 11x gains… with opportunities for improvement(!).

Page 19: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

19

From 1 to many layers…

10/20/11

Page 20: Breaking the Single-Path Barrier

Jack Baskin SoE Research Review Day

20

Future Work

• Routing protocols (path computation)– Link state– Distance vector

• Congestion management (path selection)– Routing– Network feedback

10/20/11

Page 21: Breaking the Single-Path Barrier

Thank you!