a visual exploration process for the analysis of internet routing data soon tee teoh kwan-liu ma s....

31
A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Upload: shonda-miller

Post on 03-Jan-2016

215 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

A Visual Exploration Process for the Analysis of Internet Routing Data

Soon Tee Teoh

Kwan-Liu Ma

S. Felix Wu

Presented by Zhenzhen Yan

April. 11, 2007

Page 2: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Problem

• Protect Internet against various forms of attacks, malicious uses, and errors.

• Interactive visualization for the analysis of Internet routing data.

Page 3: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Motivation

The Internet pervades many aspects of our lives:

• Commerce -An Accenture survey has found that the majority of consumers use the Internet as part of the shopping process even if they go to stores to purchase or pick up items.

• Education• Government• Production and general information

dissemination

Page 4: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Motivation Cont’d

• The Internet is vulnerable to attacks and errors.

-Researchers Estimate Worst-Case Worm Damage at $50 Billion (ENT news).

-Affect people in other unquantifiable ways.

Page 5: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Example:

128.120.0.0/16 128.120.

IP Prefix

Page 6: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

AS

• One or more networks within a single administrative domain is referred to as an Autonomous System (AS), and is assigned a unique AS number.

Page 7: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

BGP

• Border Gateway Protocol (BGP) is used to exchange network reachability information so that routers can eventually forward data packets to the correct destination.

Page 8: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

BGP announcement

Example:

128.120.0.0/16: (7,23,92)

AS-7 AS-23 AS-92 (Origin)

128.120.0.0/16: WD (no longer reachable)

Page 9: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

The Data

• The Oregon Route Views server

• 2000 and 2001

• Observation point

Page 10: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 11: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 12: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Approach

1. Aggregate Data Browsing Module

“Are there any patterns or clusters in the timing of the announcements?”

Page 13: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 14: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 15: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

2. Route Announcement Module

Page 16: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 17: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 18: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 19: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

What is the time-interval threshold to determine whether we consider a sequence of announcements part of the same “event”?

Page 20: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 21: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

3. Node/Link Visualization Module

Page 22: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 23: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 24: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 25: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

4. Peer-Comparison Module

Page 26: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 27: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

5. Events Visualization module

Page 28: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007
Page 29: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Evaluation

• The user has been able to detect instances of known types of instability events as well as identify new types, and further classify known types.

• The user was able to quickly identify faster paths, preferred paths, slow points, points of failure, as well as hypothesize about the sequence of events that led to the observations. The user also discovered new sub-classes of instability events through visualization.

• Very positive feedback from the network analysts and Internet researchers who have seen and used the visualization software.

Page 30: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Conclusions

• Two-level (overview+detail) system.

• Visualize text with timestamps.

• Arcs visualization.

• EventShrubs.

Page 31: A Visual Exploration Process for the Analysis of Internet Routing Data Soon Tee Teoh Kwan-Liu Ma S. Felix Wu Presented by Zhenzhen Yan April. 11, 2007

Questions• Why do the EventShrubs of Terralycos

and Yahoo (Microsoft and Realnetworks) show similarity?

• What will you do to better evaluate these methods?

• What other applications are likely to use these interactive visualization techniques?