Computing and Communications 1. Introduction Ying Cui Department - - PowerPoint PPT Presentation

SLIDE 1

1896 1920 1987 2006

Computing and Communications

• 1. Introduction

Ying Cui Department of Electronic Engineering Shanghai Jiao Tong University, China 2018, Autumn

1

SLIDE 2

COURSE INFORMATION

2

SLIDE 3

Lecture

• Time: Monday 8:00-9:40am, Sep 10-Dec 24 (Week 1-

16)

• Venue: Zhong 107
• Instructor: Prof. Ying Cui, IWCT, Dept. of EE

– webpage: http://iwct.sjtu.edu.cn/personal/yingcui/ – email: cuiying@sjtu.edu.cn – office: SEIEE Building 5-301A

• No textbook, research papers as references

3

SLIDE 4

Outline

• Information theory (1948)
• Coding theory (1949)
• Network coding (2000)
• Optimization
• Wireless caching (2014)
• Mobile edge computing (2015)
• Wireless multimedia transmission

4

SLIDE 5

Requirements and Grading

• Presentation (40%)

– present one paper – about 12 mins/person – last 4 weeks

• Report (60%)

– 3-page, double-column report (IEEE conference style, latex) – a review of >=5 papers in a related field and something interesting beyond the existing literature

• e.g., a comparison of different approaches in different papers, a

new problem formulation and/or solution

5

SLIDE 6

Goal

• Enrich knowledge of classic and new theories and

technologies in the area of wireless communications

• Understand how computing and communications

jointly improve performance of wireless networks

• Develop skills needed to read and write research

papers

6

SLIDE 7

COURSE OVERVIEW

7

SLIDE 8

Information Theory

• In early 1940s, it was thought impossible to send

information at a positive rate with negligible error probability over a noisy channel

• In 1948, Claude Shannon surprised the community in

[Shannon1948]

– error probability can be made nearly zero for all communication rates below channel capacity

• What is ultimate transmission rate of communication?

– channel capacity

8

[Shannon1948] C. Shannon, “A mathematical theory of communication,” Bell System Technical Journal, 1948.

SLIDE 9

Father of Information Theory

• -Claude Shannon (1916-2001)
• Found information theory with a landmark paper

[Shannon1948], in 1948 (at age of 32)

• Found digital circuit design theory in his master

thesis at MIT, in 1937 (at age of 21)

• Contribute to the field of cryptanalysis for national

defense during Word War II (by age of 29)

9

Stata Center Shannon’s Statue MIT

SLIDE 10

Coding Theory

• How to achieve channel capacity?

– channel coding (forward error correction)

• Introduce redundancy for controlling errors in data

transmission over a noisy channel

• Coding theory has been developed during the long

search for simple good codes since Shannon’s

• riginal paper in 1948

10

SLIDE 11

Network Coding

• Before advent of network coding, intermediate nodes
• nly forward incoming data flows

– independent data flows are kept separate

• Around 2000, R. Yeung et al. proposed network coding

– intermediate nodes not only forward but also process (combine) incoming independent data flows – destination nodes decode desired data flows from receiving combined data flows – combining independent flows better tailors network traffic to network environment

• Increase network throughput

11

[Yeung2000] R. Ahiswede, R. Yeung, N. Cai, S. Li and R. Yeung, “Network information flow ,” IEEE Trans. Inf. Theory,

• Apr. 2000.

SLIDE 12

Optimization

• Optimization is an abstraction of the problem of making

the best possible choice from a set of candidate choices

– portfolio optimization, device sizing in electronic circuits, data fitting

• Optimization has become an important tool in many

areas

– electronic design automation, automatic control systems, and

• ptimal design problems arising in civil, chemical, mechanical,

and aerospace engineering

• The general optimization problem is very difficult to solve
• A few problem classes can be solved reliably and

efficiently

12

[Boyd2004] Boyd and L. Vandenberghe, Convex optimization, Cambridge university press, 2004.

SLIDE 13

Wireless Caching

• Shift of wireless communication services

– connection-oriented to content-oriented services

• Name content (named data object, NDO)
• Cache popular contents at wireless edge

– caching at BSs: femto caching by Caire et al. [Carie2013] – caching at end users: coded caching by Ali et al. [Ali2014]

• Reduce delay, alleviate backhaul burden and load of

wireless links

13

[Caire2013] K. Shanmugam, N. Golrezaei, A. Dimakis, A. Molish and G. Caire, “FemtoCaching: wireless video content delivery through distributed caching helpers,” IEEE Trans. Inf. Theory, Dec. 2013. [Ali2014] M. A. Maddah-Ali and U. Niesen, “Fundamental limits of caching,” IEEE Trans. Inf. Theory, May 2014.

SLIDE 14

Mobile Edge Computing (MEC)

• Computation-intensive and latency-sensitive

applications are emerging [Hu2015]

– on-device cameras and embedded sensors

• Enable cloud computing capabilities and an IT service

environment at the edge of the cellular network

• Reduce congestion and improve user experience

14

Navigation Virtual Reality Augmented Reality

[Hu2015] Y. C. Hu, M. Patel, D. Sabella, N. Sprecher, and V. Young, “Mobile edge computing - a key technology towards 5g,” ETSI White Paper, vol. 11, 2015.

SLIDE 15 User 1

2.0 Multicast 3.0 1.0 1.0 2.5 4.0 1.0 2.0 3.0 4.0 Original view Synthesized view

Scene

1.0 2.0 3.0 4.0 1.5 2.5 3.5 4.0

1

1.0 r

2

2.0 r

3

3.0 r

4

4.0 r

User 2 User 3 User 4

1.0 1.0 2.5 2.5 4.0 4.0

Sever

Wireless Multimedia Transmission

• Video providing immersive viewing experience

– multi-view video (MVV)

• capture from different angles with multiple cameras
• user can freely select among multiple view angles

– 360 virtual reality (VR) video

• capture from all directions with omnidirectional camera
• user can freely watch the scene of interest in any viewing direction

– large size, heavy burden to wireless networks

• Improve transmission efficiency

15

unicast multicast tiled 360 VR video

FoVs

that may be watched

1 2 3 4 5 6 7 8 4 3 2 1

nv nh

user2

current FoV

viewing direction user1

current FoV

user3

current FoV c

SLIDE 16

BACKGROUND AND MOTIVATION

16

SLIDE 17

Evolution of Mobile Commun. Systems

17

1G

(1980)

2G

(1990)

3G

(2000)

4G

(2010)

5G

(2020)

Analog FDMA Digital TDMA (GSM) CDMA W-CDMA CDMA2000 TD-SCDMA OFDMA SC-FDMA MIMO Massive MIMO D2D \ M2M Spectrum sharing

Voice only Text msg Picture msg Web, Multimedia Mobile TV, GPS Video on demand IP telephony Gaming HD mobile TV Video conferencing Smart house Automated driving IoT, AR, VR

100 Mbps (DL) 50 Mbps (UL)

SLIDE 18

Main Drivers: Mobile Internet and IoT

Mobile Data Traffic: Mobile Internet & IoT Connections: Thousands of time growth Up to 100 billion

18

SLIDE 19

Vision of 5G Life

• Fiber-like access data rate
• “Zero” latency user

experience

• Up to 100 million

connections/km^2

• Consistent experience

under diverse scenarios

• Smart optimization based
• n services and users

sensing

• 100 times reduction in

energy and cost per bit

19

SLIDE 20

5G Key Capabilities: The 5G Flower

• Performance

Requirements

• Efficiency

Requirements

20

SLIDE 21

5G Technology Directions

21

Massive MIMO Novel Multiple Access Full Duplex

DAC ADC ... ... 接收机射 频单元 业务 基带单元 1 2 N T M 1 2 N R M f0 f0 f0 近端 DAC ADC ... ... 接收机射 频单元 基带单元 业务 1 2 F T M 1 2 F R M f0 远端 发射机射 频单元 发射机射 频单元

Ultra-dense networking M2M D2D

SLIDE 22

5G Challenges

• Problem of information transmission with

exponential growth can not be solved in a single dimension

– computing – caching

22

SLIDE 23

Computing and Communications

23

“computation is communication limited and communication is computation limited”

• -Prof. T. Cover, Stanford Univ.

通信 （香农定律） 计算 （摩尔定律）

计算能力 通信性能

SLIDE 24

3C—Communications, Computing and Caching

24

Communications Computing Caching

SLIDE 25

EXAMPLE 1: NETWORK CODING

25

SLIDE 26

Information Exchange

• Node A transmits x1 to Node C via Relay B and Node

C transmits x2 to Node A via Relay B

• Network coding approach uses one transmission less

26

SLIDE 27

EXAMPLE 2: CODED CACHING

27

SLIDE 28

Content Delivery with User Caching

28

SLIDE 29

Traditional Uncoded Caching Scheme

2, 2, 1 K N M = = =

server shared link user requests user caches

1 2

, ,

,

D u D u

W W

1,c

W

1,u

W

2,c

W

2,u

W

1,c

W

2,c

W

1,c

W

2,c

W

1

D

W

2

D

W

29

worst-case: are different

1 2

, D D

worst-case load=1/2*2=1

SLIDE 30

Coded Caching Scheme [Ali2014]

2, 2, 1 K N M = = =

server shared link user requests user caches

1 2

,{2} { } 1 , D D

W W Å

1,{1}

W

1,{2}

W

2,{1}

W

2,{2}

W

} 1,{1

W

} 2,{1

W

} 1,{2

W

} 2,{2

W

1

D

W

2

D

W

30

worst-case load=1/2*1=1/2 worst-case: are different

1 2

, D D [Ali2014] M. A. Maddah-Ali and U. Niesen, “Fundamental limits of caching,” IEEE Trans. Inf. Theory, May 2014.

'

' , { '}

k

k D k

W

Î

Å

S S

/ 1 2 ,| + | KM N Í = = S K S

SLIDE 31

3, 3, 1 K N M = = =

1

D

W

2

D

W

server shared link user requests caches

Traditional Uncoded Caching Scheme

3

D

W

c 1,

W

u 1,

W

c 2,

W

u 2,

W

c 3,

W

u 3,

W

c 1,

W

c 2,

W

c 3,

W

c 1,

W

c 2,

W

c 3,

W

c 1,

W

c 2,

W

c 3,

W

1 2 3

, ,u u , u,

,

D D D

W W W

3 1 / M N = 3 / 2 M N =

cached uncached

worst-case: are different

1 2 3

, , D D D

31

worst-case load=2/3*3=2

SLIDE 32

3, 3, 1 K N M = = =

server shared link user requests caches

Coded Caching Scheme [Ali2014]

1 2 1 3 2 3

,{2} ,{ } ,{3} ,{1} ,{3} ,{2} 1 ,

,

D D D D D D

W W W W W W Å Å Å

1,{1}

W

1,{2}

W

1,{3}

W

2,{1}

W

2,{2}

W

2,{3}

W

3,{1}

W

3,{2}

W

3,{3}

W

} 1,{1

W

} 2,{1

W

} 3,{1

W

} 1,{2

W

} 2,{2

W

} 3,{2

W

} 1,{3

W

} 2,{3

W

} 3,{3

W

subfiles

3 3 / 1 K N K M æ ö æ ö = = ç ÷ ç ÷ è ø è ø

1

D

W

2

D

W

3

D

W

'

' , { '}

k

k D k

W

Î

Å

S S

/ 1 2 ,| + | KM N Í = = S K S

32

worst-case load=1/3*3=1

[Ali2014] M. A. Maddah-Ali and U. Niesen, “Fundamental limits of caching,” IEEE Trans. Inf. Theory, May 2014.

SLIDE 33

Load Comparison

• Worst-case load of traditional uncoded caching
• Worst-case load of coded caching

33

( ) (1 / )

U

R M K M N =

• local caching gain

load without caching normalized local cache size

relevant if local cache size

• n order of # of files

C(

) (1 / )1 / R M K M N KM N =

• +

１

local caching gain load without caching normalized global cache size global caching gain

relevant if global cache size on order of # of files coded multicasting gain available simultaneously for all requests

SLIDE 34

EXAMPLE 3: MOBILE EDGE COMPUTING

34

SLIDE 35

Navigation

• Monitor and control the movement of a craft or

vehicle from one place to another

• Four general categories

– land navigation – marine navigation – aeronautic navigation – space navigation

• Most popular navigation systems:

– Global Positioning System (GPS) – BeiDou Navigation Satellite System (BDS)

35

SLIDE 36

Computations in Navigation

• Obtain location information

– obtain accurate locations of multiple users at the same time

• Plan route

– integrate a series of factors to better plan a path

• Process panoramic images

– process a series of images due to forward, backward and

• ther operations
• High requirements for computation capability and

computation power

36

SLIDE 37

Augmented Reality (AR)

• A live direct or indirect view of a physical, real-world

environment whose elements are "augmented" by computer-generated sensory input such as sound, video, graphics or GPS data

37

SLIDE 38

Augmented Reality (AR)

• Five critical components in an AR application:

– a video source

• obtain raw video frames from mobile camera

– a tracker

• track user position

– a mapper

• build environment model

– an object recognizer

• identify known objects

– a render

• prepare processed frame for display

38

SLIDE 39

cuiying@sjtu.edu.cn iwct.sjtu.edu.cn/Personal/yingcui

39