The document discusses research on geographic routing in social networks. It examines routing strategies using only local information on networks modeled from real-world call detail records. The strategies include depth-first search (DFS), selecting the geographically closest friend (GEO), and selecting the best connected friend (DEG). Experiments show these strategies can successfully route messages between random pairs in the networks, with performance varying based on network structure and error rate. Shorter path lengths are found when routing at the municipal level compared to the province level.
6. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
7. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
8. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
9. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
10. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
11. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
12. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
13. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
14. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
15. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
16. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
17. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
18. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
19. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
20. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
21. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
22. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
23. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
24. What’s been done so far?
network
models
Real-world
experiments
Simulation
on network
data
25. Take home message
Routing in a real world network using
only local information may be possible:
We need to understand the role of
geography in the social links
26. Pretty BIG Data
3 phone networks country scope during
6 month
Mutual links considered (at least one
interaction per direction)
User are geo-located to their most used
tower
27. Pretty BIG Data 1
0.1
0.01
network
0.001
France
P(k)
1e−04
Portugal
1e−05 Spain
1e−06
1e−07
1e−08
1 10 100 1000 10000
k
30. Let’s route!
“City”=province (20 Portugal, 96
France, 52 Spain)
We try to deliver the message to the
city
Experiment conditions:
• 60K Random pairs
• Error if not delivered in 100
hops
31. Let’s route!
“City”=province (20 Portugal, 96
France, 52 Spain)
We try to deliver the message to the
city
Experiment conditions:
• 60K Random pairs
• Error if not delivered in 100
hops
32. Let’s route!
Proposed methods
DFS: we don’t send to others who already got
the message, in case everybody got, send back
to the first person who sent us the message
GEO: send to the friend geographically closest
to target
DEG: send to the best connected friend
36. Let’s route!
“City”=municipality (305
Portugal, 3520 France, 8410
Spain)
We try to deliver the message to the
city
Experiment conditions:
• 60K Random pairs
• Error if not delivered in 100
hops