information distribution on a bus based opportunistic
play

Information distribution on a bus-based opportunistic network - PowerPoint PPT Presentation

Politecnico di Torino Information distribution on a bus-based opportunistic network Candidate Supervisor Claudio Fiandrino Paolo Giaccone November 26, 2012 Title analysis Information distribution Title analysis Comparison among routing


  1. Politecnico di Torino Information distribution on a bus-based opportunistic network Candidate Supervisor Claudio Fiandrino Paolo Giaccone November 26, 2012

  2. Title analysis Information distribution

  3. Title analysis Comparison among routing strategies: flood- Information ing and social-aware forwarding strategies distribution

  4. Title analysis Comparison among routing strategies: flood- Information ing and social-aware forwarding strategies distribution bus-based

  5. Title analysis Comparison among routing strategies: flood- Information ing and social-aware forwarding strategies distribution The backbone of the network is realized by buses: the bus schedule helps in develop- bus-based ing in a simple manner a mobility model

  6. Title analysis Comparison among routing strategies: flood- Information ing and social-aware forwarding strategies distribution The backbone of the network is realized by buses: the bus schedule helps in develop- bus-based ing in a simple manner a mobility model opportunistic network

  7. Title analysis Comparison among routing strategies: flood- Information ing and social-aware forwarding strategies distribution The backbone of the network is realized by buses: the bus schedule helps in develop- bus-based ing in a simple manner a mobility model The architecture is a kind of Delay Tolerant opportunistic Networks in which each node acts as a relay network

  8. Title analysis Comparison among routing strategies: flood- Information ing and social-aware forwarding strategies distribution The backbone of the network is realized by buses: the bus schedule helps in develop- bus-based ing in a simple manner a mobility model The architecture is a kind of Delay Tolerant opportunistic Networks in which each node acts as a relay network 2 of 19

  9. Outline The architecture 1 Mobility Model 2 Information Distribution 3 Flooding Social-aware routing algorithms 3 of 19

  10. The reference architecture Delay Tolerant Networks (DTNs) are composed of independent regions connected by gateways. When each node acts as a DTN gateway DTNs are also called Opportunistic Networks. 4 of 19

  11. Outline The architecture 1 Mobility Model 2 Information Distribution 3 Flooding Social-aware routing algorithms 5 of 19

  12. The mobility model Human mobility models are very difficult to be predicted. Google Transit Feed provides public bus schedule information. Parameters of the mobility model Torino Google Transit Feed Data; relevance r : is the number of bus passages per stop; uniformity coefficient α : describes the relation between passenger deployment and relevance. Passengers move according to p up p down 6 of 19

  13. The parameters of the mobility model Uniformity coefficient: � 0 passengers deployed proportionally to the stop relevance; α = 1 passengers deployed independently of the stop relevance. Relevance: r i = r i · ( 1 − α ) + ( α r max ) ˜ where r max = max { r i } The probability to get off the bus: r i p down = � n j = i r j The probability to get on the bus: p up = 1 − p down 7 of 19

  14. The parameters of the mobility model Uniformity coefficient: � 0 passengers deployed proportionally to the stop relevance; α = 1 passengers deployed independently of the stop relevance. Relevance: r i = r i · ( 1 − α ) + ( α r max ) ˜ where r max = max { r i } The probability to get off the bus: r i p down = � n j = i r j The probability to get on the bus: p up = 1 − p down 7 of 19

  15. The parameters of the mobility model Uniformity coefficient: � 0 passengers deployed proportionally to the stop relevance; α = 1 passengers deployed independently of the stop relevance. Relevance: r i = r i · ( 1 − α ) + ( α r max ) ˜ where r max = max { r i } The probability to get off the bus: r i p down = � n j = i r j The probability to get on the bus: p up = 1 − p down 7 of 19

  16. Map with the relevance of the stops Highest relevance 0 10 20 Lowest relevance km 8 of 19

  17. Outline The architecture 1 Mobility Model 2 Information Distribution 3 Flooding Social-aware routing algorithms 9 of 19

  18. The target Proximity-based communications. Compare the performances of: flooding; social-aware algorithms. Flooding simple; the cost in terms of network resources utilization is high. Social-aware algorithms require a priori human relation knowledge; are less aggressive in consume network resources; lead anyway to good performances. 10 of 19

  19. Outline The architecture 1 Mobility Model 2 Information Distribution 3 Flooding Social-aware routing algorithms 11 of 19

  20. Flooding: evaluation conditions Evaluation of: stop infection process ; passengers data diffusion ; Content injection in: peripheral stop; medium-relevant stop; hub stop. Different initial passenger deployment. The population consists of 100 000 passengers. The simulation period is 8:00-12:00 am. 12 of 19

  21. Flooding: performances Stop infection process · 10 5 1 3000 0 . 9 2750 0 . 8 2500 Num. Users infected 0 . 7 2250 2000 0 . 6 Num. Stops 1750 0 . 5 1500 0 . 4 1250 hub α = 0 1000 0 . 3 medium-rel α = 0 750 0 . 2 peripheral α = 0 500 hub α = 1 0 . 1 hub node 250 medium-rel α = 1 medium-rel node 0 0 peripheral α = 1 peripheral node 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 8 0 8 0 8 0 8 0 8 0 9 0 9 0 9 0 9 0 9 0 0 1 0 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 1 1 2 1 : 8 : 8 : 8 : 8 : 8 : 9 : 9 : 9 : 9 : 9 : 0 : 0 : 0 : 0 : 0 : 1 : 1 : 1 : 1 : 1 : 2 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 : : : : : : : : : : : : : : : : : : : : : 0 0 2 1 4 2 6 3 8 4 0 0 2 1 4 2 6 3 8 4 0 0 2 1 4 2 6 3 8 4 0 0 2 1 4 2 6 3 8 4 0 0 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 Time Time 13 of 19

  22. Flooding: performances Passenger data diffusion process · 10 5 1 3000 0 . 9 2750 0 . 8 2500 Num. Users infected 0 . 7 2250 2000 0 . 6 Num. Stops 1750 0 . 5 1500 0 . 4 1250 hub α = 0 1000 0 . 3 medium-rel α = 0 750 0 . 2 peripheral α = 0 500 hub α = 1 0 . 1 hub node 250 medium-rel α = 1 medium-rel node 0 0 peripheral α = 1 peripheral node 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 8 0 8 0 8 0 8 0 8 0 9 0 9 0 9 0 9 0 9 0 0 1 0 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 1 1 2 1 : 8 : 8 : 8 : 8 : 8 : 9 : 9 : 9 : 9 : 9 : 0 : 0 : 0 : 0 : 0 : 1 : 1 : 1 : 1 : 1 : 2 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 : : : : : : : : : : : : : : : : : : : : : 0 0 2 1 4 2 6 3 8 4 0 0 2 1 4 2 6 3 8 4 0 0 2 1 4 2 6 3 8 4 0 0 2 1 4 2 6 3 8 4 0 0 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 Time Time 13 of 19

  23. Outline The architecture 1 Mobility Model 2 Information Distribution 3 Flooding Social-aware routing algorithms 14 of 19

  24. Social model Model based on the concept of social space : mono-dimensional [ 0 , 1 ] ; user mapping based on the degree of interest in the content; forwarding when the social distance is below the infection radius R ; an example: 0 1 u 1 u 2 u 3 u 4 u 5 u 6 R R 15 of 19

  25. Social-aware forwarding schemes Deterministic forwarding scheme (DFS): passengers are always altruistic. d ( A , B ) < R Probabilistic forwarding scheme (PFS): content forwarded likely to social-neighbours. P ( A communicate with B ) = 1 − d ( A , B ) 2 R DFS PFS p forwarding p forwarding 1 1 0 0 R social distance 2 R social distance 16 of 19

  26. Social model: performance evaluation Analysis have been performed: in a multi-hop fashion (whole population, several timeslots); in a single-hop fashion (limited population, one timeslot); considering: a social-oblivious mobility model (SOM); a social-based mobility model (SBM). 17 of 19

  27. Social model: performance evaluation Analysis have been performed: in a multi-hop fashion (whole population, several timeslots); in a single-hop fashion (limited population, one timeslot); considering: a social-oblivious mobility model (SOM); a social-based mobility model (SBM). Results proved that in: multi-hop analysis: PFS - DFS in both mobility models; single-hop analysis: PFS - DFS in SOM; DFS - PFS in SBM; 17 of 19

  28. Social model: performance evaluation Analysis have been performed: in a multi-hop fashion (whole population, several timeslots); in a single-hop fashion (limited population, one timeslot); considering: a social-oblivious mobility model (SOM); a social-based mobility model (SBM). Results proved that in: multi-hop analysis: PFS - DFS in both mobility models; single-hop analysis: PFS - DFS in SOM; DFS - PFS in SBM; Selected scheme Comparison between flooding and DFS 17 of 19

Recommend


More recommend