26
SANS A Simple Ad hoc Network Simulator Nicolas Burri Roger Wattenhofer Yves Weber Aaron Zollinger

SANS A Simple Ad hoc Network Simulator

Embed Size (px)

DESCRIPTION

SANS A Simple Ad hoc Network Simulator. Nicolas Burri Roger Wattenhofer Yves Weber Aaron Zollinger. Bob. Alice. WLAN at Home (Infrastructure Mode). Bob. Alice. WLAN in the Woods (Ad Hoc Mode). Multihop routing. ?. Routing. G=(V,E). t. s. Routing: Internet vs. Ad Hoc Networks. - PowerPoint PPT Presentation

Citation preview

Page 1: SANS A Simple Ad hoc Network Simulator

SANSA Simple Ad hoc Network Simulator

Nicolas BurriRoger Wattenhofer

Yves WeberAaron Zollinger

Page 2: SANS A Simple Ad hoc Network Simulator

2

WLAN at Home (Infrastructure Mode)

AliceBob

Page 3: SANS A Simple Ad hoc Network Simulator

3

WLAN in the Woods (Ad Hoc Mode)

AliceBob

Page 4: SANS A Simple Ad hoc Network Simulator

4

Routing

G=(V,E)

s

t

?

Multihop routing

Page 5: SANS A Simple Ad hoc Network Simulator

5

Routing: Internet vs. Ad Hoc Networks

Internet Ad Hoc Networks

• Dedicated routers • No (or little) pre-deployed infrastructure

• Mostly static structure • Highly dynamic topology

• Low packet loss • High packet loss

• Unlimited energy • Battery lifetime

Page 6: SANS A Simple Ad hoc Network Simulator

6

Consequences

• Routing and Medium Access Control (MAC) are difficult tasks in ad hoc networks– Existing solutions for wired networks are not well suited for use in

ad hoc networks

– New paradigms and algorithms need to be developed

Lots of work for coming generations of computer scientists

But how should we teach them?

Page 7: SANS A Simple Ad hoc Network Simulator

7

Teaching the Concepts of Ad Hoc Networks

Theory

Graph Theory

Existing Algorithms Electrical Engineering

Existing Systems

Practical Exercises

Implementation of Applications

Page 8: SANS A Simple Ad hoc Network Simulator

8

Practical Exercises

• Requirements– One or more devices offering

a radio network interface• Notebook• PDA

– Test network consisting of several nodes

Page 9: SANS A Simple Ad hoc Network Simulator

9

Students as Nodes

• Each student represents a node of the graph

– Nearly no control over the topology

– Requires many participants and a lot of space

– Results of experiments are not reproducible

– Debugging is almost impossible

• Is there a better solution?

Simulation is an alternative

Page 10: SANS A Simple Ad hoc Network Simulator

10

Simulation

• Each instance of the program represents a virtual node• The simulating PC controls the network topology

Page 11: SANS A Simple Ad hoc Network Simulator

11

Existing Simulators

• Various simulators exist focusing – on functionality

– not on usability for untrained users

• ns2 General purpose simulator for all network layers Very powerful

Requires special scripts Highly complex to use

Page 12: SANS A Simple Ad hoc Network Simulator

12

SANS

• SANS has been designed for use in exercises

Intuitive “Point and Click” user interface Support for generic Java programs Platform independent Real-time simulation execution Programs developed in SANS also run on real hardware Small size (70 kB)

Limited underlying communication protocol Limited scalability

Page 13: SANS A Simple Ad hoc Network Simulator

13

SANS: Interface

Page 14: SANS A Simple Ad hoc Network Simulator

14

SANS: Interface

Page 15: SANS A Simple Ad hoc Network Simulator

16

SANS: Interface

Page 16: SANS A Simple Ad hoc Network Simulator

17

SANS: Adding Edges

Page 17: SANS A Simple Ad hoc Network Simulator

18

SANS: Link Properties

Transmission delay

Packet drop policy

Link direction

Page 18: SANS A Simple Ad hoc Network Simulator

19

SANS: Flooding Example

Page 19: SANS A Simple Ad hoc Network Simulator

20

SANS: Flooding Example

Page 20: SANS A Simple Ad hoc Network Simulator

21

SANS: Flooding Example

Page 21: SANS A Simple Ad hoc Network Simulator

22

SANS: Flooding Example

Page 22: SANS A Simple Ad hoc Network Simulator

23

SANS: Flooding Example

Page 23: SANS A Simple Ad hoc Network Simulator

24

SANS: Console Output

Page 24: SANS A Simple Ad hoc Network Simulator

25

SANS in use

• “Mobile Computing” at ETH Zurich• Server-less instant messenger as a long term homework

– Find other users in the network

– Send and receive messages to users not within direct communication range

– Relay messages for other users

• The same task was assigned in three consecutive years

Page 25: SANS A Simple Ad hoc Network Simulator

26

SANS in use

• The number of students was about 90 people and did not fluctuate much between 2002 and 2004

• 2-3 students/team• Number of handed-in complete solutions by year:

1 2

27

0

5

10

15

20

25

30

Handed-in solutions

2002 2003 2004Year

Not only due to

SANS

Page 26: SANS A Simple Ad hoc Network Simulator

27

The End

Download SANS at

http://dcg.ethz.ch/projects/SANS/Simulator.jar