CACHE CONTENT PLACEMENT USING TRIANGULAR NETWORK CODING Pouya - - PowerPoint PPT Presentation

cache content placement
SMART_READER_LITE
LIVE PREVIEW

CACHE CONTENT PLACEMENT USING TRIANGULAR NETWORK CODING Pouya - - PowerPoint PPT Presentation

1 CACHE CONTENT PLACEMENT USING TRIANGULAR NETWORK CODING Pouya Ostovari, Abdallah Khreishah, and Jie Wu Computer & Information Sciences Department, Temple University, USA Center for Networked Computing Agenda 2 Introduction


slide-1
SLIDE 1

CACHE CONTENT PLACEMENT USING TRIANGULAR NETWORK CODING

Pouya Ostovari, Abdallah Khreishah, and Jie Wu Computer & Information Sciences Department, Temple University, USA

Center for Networked Computing

1

slide-2
SLIDE 2

Agenda

 Introduction  Motivation  Content placement algorithm  Simulation  Conclusion

2

slide-3
SLIDE 3

Alice and Bob (No coding)

Alice R Bob

3

slide-4
SLIDE 4

Alice and Bob (No coding)

Alice R Bob X

3

slide-5
SLIDE 5

Alice and Bob (No coding)

Alice R Bob Y

3

slide-6
SLIDE 6

Alice and Bob (No coding)

Alice R Bob Y

3

slide-7
SLIDE 7

Alice and Bob (No coding)

Alice R Bob X

3

slide-8
SLIDE 8

Alice and Bob (No coding)

Alice R Bob X

4 transmissions

3

slide-9
SLIDE 9

Alice and Bob (Coding)

Alice R Bob

4

slide-10
SLIDE 10

Alice and Bob (Coding)

Alice R Bob X

4

slide-11
SLIDE 11

Alice and Bob (Coding)

Alice R Bob Y

4

slide-12
SLIDE 12

Alice and Bob (Coding)

Alice R Bob X+Y

4

slide-13
SLIDE 13

Alice and Bob (Coding)

Alice R Bob X+Y

3 transmissions

4

slide-14
SLIDE 14

Motivation

 Providing more amount of data to the users.

5

slide-15
SLIDE 15

Setting

 h video layers on the server:  Layer is not useful without the layers with a

smaller index.

6

slide-16
SLIDE 16

Setting

 Capacity=size of the video layers  Objective: maximizing the

total number of available layers.

7

slide-17
SLIDE 17

Triangular Coding

 Linear Coding 

ways to code h layers.

different possible placements for n caches.

 Triangular network coding  The encoded video layers are in the form .

Original packets Linear coding Triangular coding

8

slide-18
SLIDE 18

Content Placement Algorithm

 The problem of efficient content placement on the caches is an

NP-complete problem.

 The greedy algorithm fills-up the caches in rounds.  In each round, we select a user and fill-up its adjacent caches.  Selection rules  Rule 1: the user with the minimum degree.  Rule 2: the user with a larger number of filled-up caches.  Rule 3: the user whose adjacent caches have less cumulative ranks.  The algorithm fills-up the empty adjacent caches to user

with a random linear combination of the first video layers.

9

slide-19
SLIDE 19

c1

Example

 Step 1: user

has the minimum degree.

 2-0+0=2  Step 2: user has 2 filled adjacent

caches.

 3-2+2=3  Step 3: select or randomly

(assume ).

 3-2+2=3

c2 c3 c4 u3 u2 u1 u4

10

slide-20
SLIDE 20

c1

Example

 Step 1: user

has the minimum degree.

 2-0+0=2  Step 2: user has 2 filled adjacent

caches.

 3-2+2=3  Step 3: select or randomly

(assume ).

 3-2+2=3

c2 c3 c4 u3 u2 u1 u4

10

slide-21
SLIDE 21

c1

Example

 Step 1: user

has the minimum degree.

 2-0+0=2  Step 2: user has 2 filled adjacent

caches.

 3-2+2=3  Step 3: select or randomly

(assume ).

 3-2+2=3

c2 c3 c4 u3 u2 u1 u4 p1+p2 p1+p2

10

slide-22
SLIDE 22

c1

Example

 Step 1: user

has the minimum degree.

 2-0+0=2  Step 2: user has 2 filled adjacent

caches.

 3-2+2=3  Step 3: select or randomly

(assume ).

 3-2+2=3

c2 c3 c4 u3 u2 u1 u4 p1+p2 p1+p2

10

slide-23
SLIDE 23

c1

Example

 Step 1: user

has the minimum degree.

 2-0+0=2  Step 2: user has 2 filled adjacent

caches.

 3-2+2=3  Step 3: select or randomly

(assume ).

 3-2+2=3

c2 c3 c4 u3 u2 u1 u4 p1+p2 p1+p2 p1+p2+p3

10

slide-24
SLIDE 24

c1

Example

 Step 1: user

has the minimum degree.

 2-0+0=2  Step 2: user has 2 filled adjacent

caches.

 3-2+2=3  Step 3: select or randomly

(assume ).

 3-2+2=3

c2 c3 c4 u3 u2 u1 u4 p1+p2 p1+p2 p1+p2+p3

10

slide-25
SLIDE 25

c1

Example

 Step 1: user

has the minimum degree.

 2-0+0=2  Step 2: user has 2 filled adjacent

caches.

 3-2+2=3  Step 3: select or randomly

(assume ).

 3-2+2=3

c2 c3 c4 u3 u2 u1 u4 p1+p2 p1+p2 p1+p2+p3 p1+p2+p3

10

slide-26
SLIDE 26

Simulation Setting

 Simulator in the MATLAB environment.  Comparison  Number of available layers to the users.  Average utility: the number of available layers to a user divided by

its degree.

 Fairness: we define unfairness as the average difference between the

number of available layers to each user and the average number of available layers to the users.

11

slide-27
SLIDE 27

Simulations

  • Number of caches: 5
  • Number of layers: 4
  • Number of caches: 5
  • Number of layers: 4

12

slide-28
SLIDE 28

Simulations

  • Number of caches: 5
  • Number of layers: 4
  • Number of caches: 5
  • Number of layers: 4

13

slide-29
SLIDE 29

Simulations

  • Number of caches: 5
  • Number of layers: 4
  • Number of caches: 5
  • Number of layers: 4

14

slide-30
SLIDE 30

Summary

 The problem of efficient content placement.

  • n the caches is known as an NP-complete problem.

 Triangular network coding can reduce the

complexity of content placement compared to the general form of coding.

 We propose a heuristic algorithm to solve the

problem.

15

slide-31
SLIDE 31

Questions

16