Infrastructures for Virtual ISPs, Study notes of Wireless Networking

Download Infrastructures for Virtual ISPs and more Study notes Wireless Networking in PDF only on Docsity!

Encouraging Cooperation in Multi-HopWireless Networks

Ratul Mahajan, Maya Rodrig, David Wetherall

and John Zahorjan

University of Washington, June 2004.

Wireless for the masses djw // UW-CSE, 6-23-

Our goal: •^

Inexpensively extend reach Via clients, PCs: •^

Use what you find

-^

Shared infrastructure

-^

Multi-hop

-^

Ubiquitous connectivity

-^

Multiple parties

Today: •^

High-quality connectivity, butbut limited areas, high cost Via APs: •^

Carefully planned

-^

Separately provisioned

-^

Single hop, client to AP

-^

All-or-nothing access

-^

Single administration

What is the appropriate system architecture?

Outline 1. djw // UW-CSE, 6-23-

Our testbed

How to cheat, and its impact

CATCH, our solution

Some results

1. In-building 802.11 Testbed • djw // UW-CSE, 6-23-

15 nodes on one floor–

Also covered by ~10 APs

•^

Atheros, Prism II.5 based cards–

Currently one radio per PC

•^

Wired for manageability

184 feet

2. How to cheat and get away with it • djw // UW-CSE, 6-23-

Just discard unwanted packets

-^

Watchdog detects but doesn’tpunish; not a solution.

•^

Simpler, better: just don’tacknowledge connectivity

-^

Routes via cheater can’t beinferred due to asymmetry

6

4

1

2 3

5

7

8

6

4

1

2 3

5

7

8

The adverse impact of cheating (II) djw // UW-CSE, 6-23-

Even a couple of cheaters can partition high-qualitylinks, and rampant cheating ruins connectivity

3. CATCH, our solution • djw // UW-CSE, 6-23-

Goal is to detect cheaters and isolate them for a period –^

A credible threat to encourage cooperative forwarding

•^

Two difficult problems: –^

Determine when a node discards packets, even though only the nodenode knows which packets it received

-^

Get neighbors to agree to punish it, even though they must coordinatecoordinate their actions via the cheating node

•^

Approach/Insight: –^

Leverage

anonymous challenges

, where receiver doesn’t know the

the identity of the sender. Can do this with current hardware.

How to detect that a node is cheating • djw // UW-CSE, 6-23-

Combine anonymous challenges (which tell you true connectivity) withwith watchdog (below, which tells you behavior).

-^

These are statistical tests in CATCH.

6

4

1 3 2 5 7 8

Use anonymous challenges and signaling by absence

• Testbed performance – multi-hop djw // UW-CSE, 6-23-
  • Time for 6MB transfer from node 8 to nodes 4, 6, 14, &
• How all nodes can isolate a cheater • djw // UW-CSE, 6-23- - H - H - 8 H - H - H - {H - , H1 - , H1 - , H1 - } - {H - , H1 - , H1 - , H1 - }
  • {H
  • , H1
  • , H1
  • , H1
  • } - {H - , H1 - , H1 - , H1 - } - H - H - H - H - {H - ,? , H0 - , H0 - } - {H - , -? , H - , H0 - }
    • {H
    • , H0
    • , H0
    • , H0
    • } - {H - , H0 - ,? , H0 - } - {H - , H1 - , H1 - , H1 - } - {H - , H0 - ,? , H0 - }

4. Results under wireless conditions djw // UW-CSE, 6-23-

The more you cheat, the more quickly you are caught.

Rapid detection with few false positives djw // UW-CSE, 6-23-

A simulation that expands on the prior result. (An ideal

result would hug the x-y axis. Note the log scale.)

More sophisticated cheats • djw // UW-CSE, 6-23-

Drop a fraction of packets

•^

Target only some neighbors

•^

Cheat only some of the time

•^

Cheat different neighbors at different times

•^

Combinations of the above …

•^

Or, physical layer hints to undermine anonymity:^ –

Using per packet received signal strength

CATCH with signal-strength cheats djw // UW-CSE, 6-23-

Signal strength helps about half the time;

CATCH still offers some protection