Transcript Wireless Networking & Mobile Computing ECE 256, CS 215 Spring 2010 Romit Roy Choudhury Dept.

Wireless Networking & Mobile Computing
ECE 256, CS 215
Spring 2010
Romit Roy Choudhury
Dept. of ECE and CS
1
Course Logistics
2
Welcome to ECE 256
 Timings:
 Location:
 Course TA:
M/W 1:15pm to 2:30pm
125 Hudson Hall
Souvik Sen [email protected]
Office hours: TBD
 Insructor:
Romit Roy Choudhury
Ph.D from UIUC, 2006
Research in Networks, Dist Sys, Mobile Comp.
Email me at [email protected]
Visit me at 203 Hudson Hall
3
Welcome to ECE 256
 Prerequisite:
ECE 156 or CS 114
Else, come and talk to me
 Grading:




Presentation/Participation:
Homework (Paper reviews, etc.):
1 mid-term exam:
Semester-long project:
10%
20%
20%
50%
4
Welcome to ECE 256
 Class broadcast email:
[email protected]
 Course Website:
http://www.ee.duke.edu/~romit/courses/s10/ece256-sp10.html
 Most course related information will be posted on the website
 Please check course website frequently
5
Welcome to ECE 256
 Make up classes
 Will be occasionaly necessary due to travel
 Would like to schedule on a case by case basis
6
Course Theme
7
Shifting Trends
 The edge of the internet becoming wireless
 Single hop networks
 Multi-hop networks
8
Many Benefits due to Wireless
 Significantly lower cost
 No cable, low labor cost, low maintenance
 Ease
 Minimum infratructure - scatter and play
 Unrestricted mobility
 Unplugged from power outlet
 Ubiquity
 Available like water/electricity - holy grail
9
Changing Face of Mobile Devices
 Small, powerful, and sensor-embedded
 Always with us … always on
 Shifting from reactive to proactive computing
QuickTime™ and a
decompressor
are needed to see this picture.
QuickTime™ and a
decompressor
are needed to see this picture.
10
If everything goes right,
the future will be
“A concious invisible Internet hanging from
a passive, visible Internet …”
11
The Future
Internet
QuickTime™ and a
decompressor
are needed to see this picture.
12
The Future
Microsoft, Intel, Cisco …
Mesh Networks and
Wireless Backbones
QuickTime™ and a
decompressor
are needed to see this picture.
Internet
Personal Area
Networks
Motorola, Intel,
Samsung …
RFID and
Sensor Networks
Citywatchers, Walmart
Intel, Philips, Bosch …
13
But, what does it take for that
mobile/wireless future to become feasible?
14
Applications that exploit
ubiquity and mobility.
Challenges underlying such
applications
Research
Ubiquitous Services
Incentives
Application
Security
Loss Discrimination
Energy Savings
Spatial Reuse
Eavesdropping
Transport
Network
MAC / Link
PHY
Enabling wireless ubiquity.
Showing what is feasible, and
what is not …
Privacy
Mobility
Interference Mgmt.
Channel fluctuations
15
Mobile
Computing
Research
Ubiquitous Services
Incentives
Application
Security
Loss Discrimination
Energy Savings
Spatial Reuse
Eavesdropping
Transport
Network
MAC / Link
PHY
Wireless
Networking
Privacy
Mobility
Interference Mgmt.
Channel fluctuations
16
This Course
 Introduces fundamentals of wireless channel
 The departure from wired networks …
 Emerging innovations in EE, communications area
 Exposes implications on protocol design
 At MAC, Network, Transport, Security
 Investigates gap between idea and and actual system
 Considers theoretical aspects
 Envisions new mobile computing applications
 Identifies challenges underlying them
 Resolves these challenges into a full system solution
 Allows you to design/develop your own ideas
 Ideally extending the state of the art
17
At the End of this Course …
 You understand




Physical layer (radios, rate, antennas, channels)
MAC protocols (who gets the chance to talk)
Routing (path selection algorithms and issues)
Reliability (wireless congestion control, rate control)
 Applications (social networks, personal networks, P2P networks)
 Human sensing, Urban sensing
•
•
•
•
Localization (extracting the location of a device)
Mobility (how it helps and disrupts communication)
Interfaces (phones are more than communication devices)
Privacy (how to protect a user from being tracked)
 Energy-awareness (how it percolates various network functions)
 Emerging Topics (interference cancellation, multicast, rural nets)
 Capacity (what is feasible, what are performance bounds)
18
What this Course Does Not Cover
 Not a wireless communications course
 Does not cover





Modulation schemes
Transmitter/Receiver design
Signal processing and antenna design
Source coding / channel coding
Etc.
 This is course on
 Design, analysis, and implementation of protocols
and algorithms in (mobile) wireless network systems
19
Some other Thoughts
 Dilemma
1. Teach very advanced stuff for the networking pro
2. Teach from absolute scratch for the uninitiated
I will try to strike a balance
Please bear with me if materials are sometimes
too easy/difficult for YOU
20
Course Structure
21
Course Structure
 I will present most lectures and papers
 You present once in entire semester (30 minutes)
 2 students present in one class
 For every class, read 2 of assigned papers
 Write reviews for each and email TA before class
 Bring printed copy to class
 A random set of reviews will be graded :)
 Several recommended readings
 Make an effort to read them
 I understand that you cannot do so always
22
Course Structure
 1 open-book mid term, No Final Exam
 Tentative date of mid-term: End march/early April
 Semester-long class project
 In groups of 2 (max 3)
 Focus on this from early on
 Class ends with a final project poster/demo
 Submit conference-style paper
 Prize for 3 best projects
• Potentially funded by industry
23
Class Participation / Presentation,
Reading Assignment,
and
Course Project
24
Participation / Presentation
 Ask lots of questions. Period.
 I strongly encourage you to ask, disagree, debate
 Class presentation
 You present one paper (30 minutes)
 Pick an open slot (ones not marked “Romit”)
• Earlier you pick, more options you have to choose from
• Deadline is Jan 27, 2010
 Email me your choice of paper (and date)
 Don’t worry about not knowing the topic of ppt
• By that time, you will know enough
25
Reading Assignment
 Read the papers assigned for reading
 Critic / Review them carefully
 Reviews should not be more than a page
 Email reviews to TA + Bring a hard copy to class
 Random set of reviews will be graded
 I might upload selected reviews on a webpage
26
Thoughts on Reading Papers
 Know why you are reading the paper
 Reading for absorbing concepts (class assignment)
• Read fully, think, reread, ask, challenge
 Reading for excitement (deciding project topic)
• Read initial parts, don’t try to understand everything, get a feel
 Reading for problem identification
• Read the problem carefully
Most
Important
 Reading to discriminate (before finalizing project)
• Read solution, ensure your ideas different, analyse
performance
27
Course Research Projects
 Projects consist of 3 parts:
 Problem identification
 Solution design
 Performance evaluation
 Each paper you read is someone’s project
 Many papers are actually student’s class projects
 Read them critically
 Ask yourself
• Is the problem really important ? Should you care ?
• Is the solution sound ? Under what assumptions? Do you
•
have other (better) ideas ?
Is evaluation biased ? Are reults shown only in good light?
28
More on Projects
 Discuss your thoughts, ideas with me
 They need not be cooked, and can have many flaws
 Statistically, every 18 ideas lead to one decent idea
 If you like an area / direction
 Read many many related papers
 Don’t try to come up with a quick solution
 Ensure your problem is a new, real problem
 Finding the solution is typically easy
29
More on Projects
 Protocol evaluation typically requires coding
 Think what you would like to do
 Options are:
• Coding on real devices (like sensors, phones, routers)
• Coding in existing network simulators (ns2, Qualnet, etc.)
• Coding your own simulator
• Theoretical projects involve MATLAB, CPLEX, etc.
 Project ideas take time … think now and then
 Spending 3 hours for 10 days better than 10 hours for
3 days
30
More on Projects
 Find a project partner early
 Discuss reviews, papers, potential project themes
 Class project often bottlenecked by platform
 Think of the evaluation platform during project selection
 If you are not familiar with the Linux OS, it’s a bad idea
to do a project involving router-prorgamming
31
Some Closing Thoughts
 This class is about research
 Be active, ask questions, debate, and disagree
 Don’t worry too much about grades
 It does not matter as much as you think
 Read a lot - this is a hot research area
 If you are hunting for MS/PhD area, read even more
 Interact with me
 Even if you have ZERO clue of what’s going on
32
Hello!
I am ECE 256
Any Questions?
33
ECE 256 Assignment 1
Watch:
Assignment Due:
Jan 27
Group work allowed
Theater preffered, TV acceptable
34
Questions ?
35
Welcome to ECE 256
 Please fill up student survey
 Helps me in designing the course better
36
Towards Ubiquity
 Focus on wireless multihop networks
B
A
D
C
 Initial applications in military
 Emering commercial technologies and applications
• For example …
37