Using Network Coding for Dependent Data Broadcasting in a Mobile Environment

1. Using Network Coding for Dependent Data Broadcasting in a Mobile Environment Chung-Hua Chu, De-Nian Yang and Ming-Syan Chen IEEE GLOBECOM 2007 Reporter : Lzu Lin Wang

2. Outline • INTRODUCTION • PROBLEM FORMULATION • GENERATING A BROADCAST PROGRAM • EXPERIMENTAL RESULTS • CONCLUSION

3. INTRODUCTION-Network coding • Network Coding • Instead of simply forwarding data, intermediate node encodes several input packets into one output packet.

4. INTRODUCTION-Networkcoding • Linear network coding • The number of retrieved packets has to at least larger than the number of original packets, otherwise there are several linearly dependent equations • Random network coding • The probability of selecting the linearly dependent combinations of the data items can be ignored even in small field sizes. • The traditional network coding does not leverage the data items stored in mobile users to encode all data items, and thereby the unnecessary encoding of data items leads to more access delay.

5. INTRODUCTION • DOP (Demand-Oriented Encoding Problem) • The problem is to minimize the average access time with maximizing the number of the demanding users. • Determine the optimal linear combination of the data item in the multiple channels environment • Demand-Oriented Pairing (DOP) algorithm selects the data items such that the linear combination of the selected items can satisfy the maximal number of the mobile users.

7. PROBLEM FORMULATION-Notation and Definition

8. PROBLEM FORMULATION-Notation and Definition The problem is to minimize the average access time with the maximizing , where represents the number of user members in 􀁇 .

9. PROBLEM FORMULATION-Derivation of Access Time • Average access time z : data size b: Bandwidth of broadcast channel The number of remaining time slots that the interested user needs to download after waiting for cycle i cycles The number of complete broadcast cycles for a mobile user ui to retrieve the encoded data. The length of broadcast cycle

10. PROBLEM FORMULATION-Derivation of Access Time • The number of channel =2 • User i stored data is {B,K} • Query data items is {A,D} • L is equal to 3 ( ) • is equal 1 • ATT is equal to In the same time slot

11. PROBLEM FORMULATION-Derivation of Access Time

12. GENERATING A BROADCAST PROGRAM-Demand-Oriented Pairing (DOP) algorithm • DOP creates Table DT (Demanding Table) • DOP creates Demanding Graph • DOP will delete the user members in Table DT if the encoding data can be decoded correctly by the data items stored in users • Add the user members to Table DT whose query data items correspond to new stored data items.

13. GENERATING A BROADCAST PROGRAM-DOP creates Table DT (Demanding Table)

14. GENERATING A BROADCAST PROGRAM- Demanding Graph 1 5 4 2 3 4 3 3 3 2 2 4 3

15. GENERATING A BROADCAST PROGRAM- Matching 1 5 4 2 3 4 3 3 3 2 2 4 3 3 1 6 1 1

16. GENERATING A BROADCAST PROGRAM-Matching 1 5 4 2 3 4 3 3 3 2 2 4 3 3 1 6 1 1

17. GENERATING A BROADCAST PROGRAM- Algorithm DOP E1

18. EXPERIMENTAL RESULTS • Simulation Model

19. EXPERIMENTAL RESULTS

20. CONCLUSION • Author proposed a novel data broadcasting framework that adopts network coding with leveraging the data items stored in users to reduce the average access time. • The algorithm selects the data items to be encoded according to the data items stored in users to satisfy maximal number of mobile users and reduce access delay.