NC-CELL: Network Coding-based Content Distribution in Cellular - PowerPoint PPT Presentation

IEEE GLOBECOM 2014 NC-CELL: Network Coding-based Content Distribution in Cellular Networks for Cloud Applications Claudio Fiandrino University of Luxembourg Dzmitry Kliazovich Pascal Bouvry Albert Y. Zomaya University of Sydney December 11, 2014

Agenda 1 Introduction Network coding in cellular networks (NC-CELL) 2 3 Evaluation 4 Conclusion Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 1 of 10

Outline 1 Introduction Network coding in cellular networks (NC-CELL) 2 3 Evaluation 4 Conclusion Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 1 of 10

Motivation ◮ Mobile data traffic will rise up to 15 EB per month by 2018 ◮ By 2017 4.4 billion people will use mobile cloud applications ◮ $ 45 billion market ◮ Mobile cloud applications will account for 90% of all mobile data traffic by 2018 100% Non-Cloud 10 % 12 % 15 % 14 % 18 % 17 % Cloud 50% 88 % 90 % 85 % 86 % 82 % 83 % 0 2013 2014 2015 2017 2018 2016 Source: Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2013-2018 Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 2 of 10

The key idea Optimizing information delivery of flows in mobile networks with overlapping or partially overlapping content through network coding. ◮ Geographically co-located users ◮ Mobile cloud applications content ◮ Advertisement ◮ Maps ◮ Meteo ◮ Google Now Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 3 of 10

Outline 1 Introduction Network coding in cellular networks (NC-CELL) 2 3 Evaluation 4 Conclusion Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 3 of 10

The scenario Buffers Network Coding S-GW P-GW Internet Cloud MME UE E-UTRAN Evolved Packet Core LTE Network Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 4 of 10

An example UE1 UE2 eNodeB Cloud Application Request A Packet request Send content A Packet A UE1 Cache and forward A UE1 Packet A UE1 Process and store A UE1 Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 5 of 10

An example UE1 UE2 eNodeB Cloud Application Request A Packet request Send content A Packet A UE1 Cache and forward A UE1 Packet A UE1 Process and store A UE1 Request B Packet request Send content B UE2 Packet B UE2 Cache and forward B UE2 Packet B UE2 Process and store B UE2 Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 5 of 10

An example UE1 UE2 eNodeB Cloud Application Request A Packet request Send content A Packet A UE1 Cache and forward A UE1 Packet A UE1 Process and store A UE1 Request B Packet request Send content B UE2 Packet B UE2 Cache and forward B UE2 Packet B UE2 Process and store B UE2 Request B Packet request Send content B UE1 Packet B UE1 Check if B is in buffer Coding ( A ⊕ B ) UE1 , UE2 Packet ( A ⊕ B ) UE1 , UE2 Decode B using A UE1 Decode A using B UE2 Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 5 of 10

The key aspects ◮ Monitor and cache in transit traffic ◮ Identify coding opportunities Coding Opportunities eNodeBs can deliver information needed by two or more users with a single coded transmission. ◮ XOR to combine packets Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 6 of 10

Content distribution Optimal allocation for content distribution u k c 2 k , k c k , k c n , k c n − k , 1 ⊕ c n − k + 1 , 2 c 2 k − 1 , k − 1 ⊕ c 2 k , k c k − 1 , k − 1 ⊕ c k , k c k + 1 , 1 ⊕ c k + 2 , 2 c n − 1 , k − 1 ⊕ c n , k Users c 1 , 1 ⊕ c 2 , 2 ⋮ ⋱ ⋱ ⋱ ⋮ ⋮ ⋮ ⋮ ⋮ c n − k + 1 , 2 c k + 2 , 2 u 2 c 2 , 2 c k + 1 , 1 c n − k , 1 u 1 c 1 , 1 t Individual Transmission - Encoded Transmission Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 7 of 10

Outline 1 Introduction Network coding in cellular networks (NC-CELL) 2 3 Evaluation 4 Conclusion Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 7 of 10

Throughput improvement ◮ Number of transmissions at eNodeB � n if r = 0 k · ( k + ϑ ) , σ = � n � · ( k + ϑ ) + k + ( r − 1 ) , otherwise k ◮ n : common chunks ◮ k : users · 10 4 ◮ ϑ : encoded transmissions ◮ r : remainder of n / k NC-CELL Disabled 1 Num. Transmissions 0 . 8 0 . 6 0 . 4 NC-CELL Enabled 0 . 2 0 2 4 1 000 6 800 600 8 400 200 10 k n Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 8 of 10

Evaluation ◮ Coding gain η = γ σ ◮ γ : total number of chunks 7 6 k = 10 5 Coding gain η k = 8 4 k = 6 3 k = 4 2 k = 2 1 0 10 100 200 300 400 500 Num. common chunks n Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 9 of 10

Outline 1 Introduction Network coding in cellular networks (NC-CELL) 2 3 Evaluation 4 Conclusion Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 9 of 10

Conclusion ◮ Efficient content distribution for cloud applications in mobile cellular networks ◮ Network coding and caching performed at eNodeB ◮ Considerable throughput improvement Claudio Fiandrino | IEEE GLOBECOM 2014 | NC-CELL 10 of 10

Thank You! Thank You! Thank You!