Download presentation
Presentation is loading. Please wait.
1
Jon Crowcroft Pan Hui Computer Laboratory University of Cambridge
Mob Handed: Realworld Human Mobility Model driven design of forwarding algorithms for Opportunistic Networking Jon Crowcroft Pan Hui Computer Laboratory University of Cambridge February Imperial College London – talk for Erol Gelenbe’s research group
2
Outline Multiple levels human heterogeneity
Local community structures Diversity of centrality in different scales Four categories of human relationship Heterogeneous forwarding algorithms Design space RANK (centrality based forwarding) LABEL (community based forwarding) BUBBLE RAP (centrality meets community) Approximation and predictability Decentralized approximation of centrality Human predictability February Imperial College London – talk for Erol Gelenbe’s research group
3
Understanding multiple levels of heterogeneity
The first goal of this research is to move to a third generation of human mobility models, understanding heterogeneity at multiple levels of detail. February Imperial College London – talk for Erol Gelenbe’s research group
4
Social Structures Vs Network Structures
Community structures Social communities, i.e. affiliations Topological cohesive groups or modules Centralities Social hubs, celebrities and postman Betweenness, closeness, inference power centrality February Imperial College London – talk for Erol Gelenbe’s research group
5
K-clique Community Definition
Union of k-cliques reachable through a series of adjacent k-cliques [Palla et al] Adjacent k-cliques share k-1 nodes Members in a community reachable through well-connected well subsets Examples 2-clique (connected components) 3-clique (overlapping triangles) Overlapping feature Percolation threshold pc (k)= 1/[(k-1)N]^(1/(k-1)) February Imperial College London – talk for Erol Gelenbe’s research group
6
K-clique Communities in Cambridge Dataset
February Imperial College London – talk for Erol Gelenbe’s research group
7
K-clique Communities in Infocom06 Dataset
Barcelona Group Paris Group A Paris Group B Lausanne Group Paris Groups K=3 February Imperial College London – talk for Erol Gelenbe’s research group
8
K-clique Communities in Infocom06 Dataset
Barcelona Group Paris Group A Paris Group B Lausanne Group Paris Groups K=4 February Imperial College London – talk for Erol Gelenbe’s research group
9
K-clique Communities in Infocom06 Dataset
Barcelona Group (Spanish) Paris Group A (French) Paris Group B (French) Italian K=5 February Imperial College London – talk for Erol Gelenbe’s research group
10
Other Community Detection Methodologies
Betweenness [Newman04] Modularity [Newman06] Information theory[Rosvall06] February Imperial College London – talk for Erol Gelenbe’s research group
11
Centrality in Temporal Network
Large number of unlimited flooding Uniform sourced and temporal traffic distribution Number of times on shortest delay deliveries Analogue to Freeman centrality [freeman] February Imperial College London – talk for Erol Gelenbe’s research group
12
Homogenous Centrality
Reality Cambridge Infocom06 HK February Imperial College London – talk for Erol Gelenbe’s research group
13
Within Group Centrality Cambridge Dataset
Group B Group A February Imperial College London – talk for Erol Gelenbe’s research group
14
Within Group Centrality Reality Dataset
Group A Group B Group D Group C February Imperial College London – talk for Erol Gelenbe’s research group
15
Model Node Centrality Node centrality should be modelled in different levels of heterogeneity February Imperial College London – talk for Erol Gelenbe’s research group
16
Regularity and Familiarity
Correlation Coefficient = II III IV Familiarity I: Community II. Familiar Strangers III. Strangers IV. Friends February Imperial College London – talk for Erol Gelenbe’s research group
17
c: 0.6604 c: 0.5817 Infocom05 Infocom06 c: 0.8325 c: 0.7927 Reality HK
February Imperial College London – talk for Erol Gelenbe’s research group
18
Heterogeneous Forwarding
The second goal of this research is to devise efficient forwarding algorithms for PSNs which take advantage of both a priori and learned knowledge of the structure of human mobility. February Imperial College London – talk for Erol Gelenbe’s research group
19
Imperial College London – talk for Erol Gelenbe’s research group
Thank you but you are in the opposite direction! I can also carry for you! I have 100M bytes of data, who can carry for me? Give it to me, I have 1G bytes phone flash. Don’t give to me! I am running out of storage. Reach an access point. There is one in my pocket… Internet Search La Bonheme.mp3 for me Finally, it arrive… Search La Bonheme.mp3 for me Search La Bonheme.mp3 for me February Imperial College London – talk for Erol Gelenbe’s research group
20
Interaction and Forwarding
Third generation human interaction model Categories of human contact patterns Clique and community Popularity/Centrality Dual natures of mobile network Social network Physical network Benchmark Forwarding strategies Flooding, Wait, and Multiple-copy-multiple-hop (MCP) February Imperial College London – talk for Erol Gelenbe’s research group
21
Design Space Human Dimension Network Plane Explicit Social Structure
Bubble Label Human Dimension Structure in Cohesive Group Clique Label Network Plane Rank, Degree Structure in Degree February Imperial College London – talk for Erol Gelenbe’s research group
22
Greedy Ranking Algorithm (RANK)
Use pre-calculated centrality/rank Push traffic to nodes have higher rank Good performance in small and homogeneous February Imperial College London – talk for Erol Gelenbe’s research group
23
Greedy Ranking Algorithm
Hierarchical organization Hierarchical paths [Trusina et al] High percentage in most dataset February Imperial College London – talk for Erol Gelenbe’s research group
24
Problem with RANK Heterogeneous at multiple levels
Best node for the whole system may not be best node for a specific community D C D B E A February Imperial College London – talk for Erol Gelenbe’s research group
25
Problem with RANK Hop distribution and rank at dead-end for HK dataset
February Imperial College London – talk for Erol Gelenbe’s research group
26
Label Strategy (LABEL)
Priori label, e.g. affiliation Correlated interaction Forward to nodes have same label as the destination Good performance in conference mixing environment Infocom06 February Imperial College London – talk for Erol Gelenbe’s research group
27
Problem with LABEL In a less mixing environment (e.g. Reality)
A person in one group may not meet members in another group so often Wait for destination group not efficient February Imperial College London – talk for Erol Gelenbe’s research group
28
Centrality meets Community
Population divided into communities Node has a global and local ranking Global popular node like a postman, or politician in a city Local popular node like Christophe Diot in SIGCOMM BUBBLE-A BUBBLE-B February Imperial College London – talk for Erol Gelenbe’s research group
29
Global Community Sub community Sub community Destination Ranking
Subsub community Sub community Sub community Source February Imperial College London – talk for Erol Gelenbe’s research group Global Community
30
Centrality meets Community
Bubble Reality single group February Imperial College London – talk for Erol Gelenbe’s research group
31
Making Centrality Practical
How can each node know its own centrality in decentralised way? How well does past centrality predict the future? February Imperial College London – talk for Erol Gelenbe’s research group
32
Approximating Centrality
Total degree, per-6-hour degree Correlation coefficients, and February Imperial College London – talk for Erol Gelenbe’s research group
33
Approximating Centrality
DEGREE S-Window A-Window (Exponential Smoothing) February Imperial College London – talk for Erol Gelenbe’s research group
34
Predictability of Human Mobility
Three sessions of Reality dataset Two sessions using the ranking calculated from the first session Almost same performance February Imperial College London – talk for Erol Gelenbe’s research group
35
Conclusion and Future Woks
Forwarding using priori label or social structure inferred through observation Distributed k-clique building through gossiping Why per-6-hour? Weighted version of k-clique detection Third generation modeling February Imperial College London – talk for Erol Gelenbe’s research group
36
Jon.Crowcroft@cl.cam.ac.uk Imperial College London – February 15 2007
talk for Erol Gelenbe’s research group
Similar presentations
