On Exploiting Transient Social Contact Patterns for Data Forwarding in Delay-Tolerant Networks 1 Wei Gao Guohong Cao Tom La Porta Jiawei Han Presented.

Slides:



Advertisements
Similar presentations
Enabling Inter-domain DTN Communications by Networked Static Gateways Ting He*, Nikoletta Sofra, Kang-Won Lee*, and Kin K Leung * IBM Imperial College.
Advertisements

Supporting Cooperative Caching in Disruption Tolerant Networks
Fast Algorithms For Hierarchical Range Histogram Constructions
CSLI 5350G - Pervasive and Mobile Computing Week 6 - Paper Presentation “Exploiting Beacons for Scalable Broadcast Data Dissemination in VANETs” Name:
Psychology: A Modular Approach to Mind and Behavior, Tenth Edition, Dennis Coon Appendix Appendix: Behavioral Statistics.
Table of Contents Exit Appendix Behavioral Statistics.
Generated Waypoint Efficiency: The efficiency considered here is defined as follows: As can be seen from the graph, for the obstruction radius values (200,
Assessment. Schedule graph may be of help for selecting the best solution Best solution corresponds to a plateau before a high jump Solutions with very.
Forwarding Redundancy in Opportunistic Mobile Networks: Investigation and Elimination Wei Gao 1, Qinghua Li 2 and Guohong Cao 3 1 The University of Tennessee,
1 Sensor Relocation in Mobile Sensor Networks Guiling Wang, Guohong Cao, Tom La Porta, and Wensheng Zhang Department of Computer Science & Engineering.
By Libo Song and David F. Kotz Computer Science,Dartmouth College.
Spring INTRODUCTION There exists a lot of methods used for identifying high risk locations or sites that experience more crashes than one would.
Multiobjective VLSI Cell Placement Using Distributed Simulated Evolution Algorithm Sadiq M. Sait, Mustafa I. Ali, Ali Zaidi.
Probabilistic Aggregation in Distributed Networks Ling Huang, Ben Zhao, Anthony Joseph and John Kubiatowicz {hling, ravenben, adj,
Pattern Classification All materials in these slides were taken from Pattern Classification (2nd ed) by R. O. Duda, P. E. Hart and D. G. Stork, John Wiley.
Mutual Information Mathematical Biology Seminar
Beneficial Caching in Mobile Ad Hoc Networks Bin Tang, Samir Das, Himanshu Gupta Computer Science Department Stony Brook University.
On Appropriate Assumptions to Mine Data Streams: Analyses and Solutions Jing Gao† Wei Fan‡ Jiawei Han† †University of Illinois at Urbana-Champaign ‡IBM.
SOCIAL-BASED FORWARDING SCHEMES Rance Fredericksen CMPE 257 Wireless Networks.
Peer-to-Peer Based Multimedia Distribution Service Zhe Xiang, Qian Zhang, Wenwu Zhu, Zhensheng Zhang IEEE Transactions on Multimedia, Vol. 6, No. 2, April.
Efficient and Robust Computation of Resource Clusters in the Internet Efficient and Robust Computation of Resource Clusters in the Internet Chuang Liu,
On the Construction of Energy- Efficient Broadcast Tree with Hitch-hiking in Wireless Networks Source: 2004 International Performance Computing and Communications.
Distributed Cluster Repair for OceanStore Irena Nadjakova and Arindam Chakrabarti Acknowledgements: Hakim Weatherspoon John Kubiatowicz.
Efficient Estimation of Emission Probabilities in profile HMM By Virpi Ahola et al Reviewed By Alok Datar.
Novel Self-Configurable Positioning Technique for Multihop Wireless Networks Authors : Hongyi Wu Chong Wang Nian-Feng Tzeng IEEE/ACM TRANSACTIONS ON NETWORKING,
1 Emulating AQM from End Hosts Presenters: Syed Zaidi Ivor Rodrigues.
CHAPTER 6 Statistical Analysis of Experimental Data
Jana van Greunen - 228a1 Analysis of Localization Algorithms for Sensor Networks Jana van Greunen.
How to Turn on The Coding in MANETs Chris Ng, Minkyu Kim, Muriel Medard, Wonsik Kim, Una-May O’Reilly, Varun Aggarwal, Chang Wook Ahn, Michelle Effros.
Ns Simulation Final presentation Stella Pantofel Igor Berman Michael Halperin
Wei Gao Joint work with Qinghua Li, Bo Zhao and Guohong Cao Department of Computer Science and Engineering The Pennsylvania State University Multicasting.
Chapter 9 Numerical Integration Numerical Integration Application: Normal Distributions Copyright © The McGraw-Hill Companies, Inc. Permission required.
© Y. Zhu and Y. University of North Carolina at Charlotte, USA 1 Chapter 1: Social-based Routing Protocols in Opportunistic Networks Ying Zhu and.
Wei Gao1 and Qinghua Li2 1The University of Tennessee, Knoxville
Your Friends Have More Friends Than You Do: Identifying Influential Mobile Users Through Random Walks Bo Han, Aravind Srinivasan University of Maryland.
1 BitHoc: BitTorrent for wireless ad hoc networks Jointly with: Chadi Barakat Jayeoung Choi Anwar Al Hamra Thierry Turletti EPI PLANETE 28/02/2008 MAESTRO/PLANETE.
Switch-and-Navigate: Controlling Data Ferry Mobility for Delay-Bounded Messages Liang Ma*, Ting He +, Ananthram Swami §, Kang-won Lee + and Kin K. Leung*
Prediction Assisted Single-copy Routing in Underwater Delay Tolerant Networks Zheng Guo, Bing Wang and Jun-Hong Cui Computer Science & Engineering Department,
Cooperative Caching for Efficient Data Access in Disruption Tolerant Networks.
Energy Efficient Phone-to-Phone Communication Based on WiFi Hotspots in PSN En Wang 1,2, Yongjian Yang 1, and Jie Wu 2 1 Dept. of Computer Science and.
Distributed Maintenance of Cache Freshness in Opportunistic Mobile Networks Wei Gao and Guohong Cao Dept. of Computer Science and Engineering Pennsylvania.
On Exploiting Transient Contact Patterns for Data Forwarding in Delay Tolerant Networks Wei Gao and Guohong Cao Dept. of Computer Science and Engineering.
How Small Labels create Big Improvements April Chan-Myung Kim
Routing In Socially Selfish Delay Tolerant Networks Chan-Myung Kim
ENERGY-EFFICIENT FORWARDING STRATEGIES FOR GEOGRAPHIC ROUTING in LOSSY WIRELESS SENSOR NETWORKS Presented by Prasad D. Karnik.
Salah A. Aly,Moustafa Youssef, Hager S. Darwish,Mahmoud Zidan Distributed Flooding-based Storage Algorithms for Large-Scale Wireless Sensor Networks Communications,
Selfishness, Altruism and Message Spreading in Mobile Social Networks September 2012 In-Seok Kang
Exploiting Context Analysis for Combining Multiple Entity Resolution Systems -Ramu Bandaru Zhaoqi Chen Dmitri V.kalashnikov Sharad Mehrotra.
PRoPHET+: An Adaptive PRoPHET- Based Routing Protocol for Opportunistic Network Ting-Kai Huang, Chia-Keng Lee and Ling-Jyh Chen.
Zibin Zheng DR 2 : Dynamic Request Routing for Tolerating Latency Variability in Cloud Applications CLOUD 2013 Jieming Zhu, Zibin.
User-Centric Data Dissemination in Disruption Tolerant Networks Wei Gao and Guohong Cao Dept. of Computer Science and Engineering Pennsylvania State University.
Content Dissemination in Mobile Social Networks Cheng-Fu Chou.
A Sociability-Based Routing Scheme for Delay-Tolerant Networks May Chan-Myung Kim
Lecture 2d: Performance Comparison. Quality of Measurement Characteristics of a measurement tool (timer) Accuracy: Absolute difference of a measured value.
Dual-Region Location Management for Mobile Ad Hoc Networks Yinan Li, Ing-ray Chen, Ding-chau Wang Presented by Youyou Cao.
1 Utilizing Shared Vehicle Trajectories for Data Forwarding in Vehicular Networks IEEE INFOCOM MINI-CONFERENCE Fulong Xu, Shuo Gu, Jaehoon Jeong, Yu Gu,
Social-Aware Stateless Forwarding in Pocket Switched Networks Soo-Jin SHIN
PCB 3043L - General Ecology Data Analysis.
Cooperative Location- Sensing for Wireless Networks Authors : Haris Fretzagias Maria Papadopouli Presented by cychen IEEE International Conference on Pervasive.
Multicasting in delay tolerant networks a social network perspective networks October2012 In-Seok Kang
© SITILabs, University Lusófona, Portugal1 Chapter 2: Social-aware Opportunistic Routing: the New Trend 1 Waldir Moreira, 1 Paulo Mendes 1 SITILabs, University.
Load Balanced Link Reversal Routing in Mobile Wireless Ad Hoc Networks Nabhendra Bisnik, Alhussein Abouzeid ECSE Department RPI Costas Busch CSCI Department.
1 Travel Times from Mobile Sensors Ram Rajagopal, Raffi Sevlian and Pravin Varaiya University of California, Berkeley Singapore Road Traffic Control TexPoint.
Jump to first page Inferring Sample Findings to the Population and Testing for Differences.
Performance Comparison of Ad Hoc Network Routing Protocols Presented by Venkata Suresh Tamminiedi Computer Science Department Georgia State University.
2010 IEEE Global Telecommunications Conference, GLOBECOM 2010 Zheng Guo; Bing Wang; Jun-Hong Cui Prediction Assisted Single-copy Routing in Underwater.
Mingze Zhang, Mun Choon Chan and A. L. Ananda School of Computing
1 Exploiting Friendship Relations for Efficient Routing in Mobile Social Networks Eyuphan Bulut, Student Member, IEEE, and Boleslaw K Szymanski, Fellow,
PCB 3043L - General Ecology Data Analysis.
Gerald Dyer, Jr., MPH October 20, 2016
Presentation transcript:

On Exploiting Transient Social Contact Patterns for Data Forwarding in Delay-Tolerant Networks 1 Wei Gao Guohong Cao Tom La Porta Jiawei Han Presented by Michael Conlan

Agenda Introduction Overview of Network Model and Algorithm Trace-Based Pattern Formulation Data Forwarding Metric Exploiting Transient Community Structure Performance Evaluation Conclusion

Introduction: Problem Statement  Delay Tolerant Networks (DTN) populated by mobile devices have intermittent connectivity and low node density  Data forwarding metrics determined by stochastic processes and predicted node mobility limited by human randomness  Problem Statement: How best to forward/relay data in DTNs to ensure timely and efficient delivery?

Introduction: Social Contact Patterns  Node forwarding capability characterized by their Social Contact Patterns:  Centrality – connectivity to many nodes that enables wider or faster delivery  Community – naturally occurring grouping of connected nodes  Consider on global scope and local scope  Most social aware forwarding schemes based on cumulative social contact patterns  BUT cumulative contact patterns differ from transient contact patterns

Introduction: Proposed Solution  Proposed Solution: Exploit Transient Social Contact Patterns to improve data forwarding by considering these perspectives of contact patterns:  Transient Contact Distribution – rate of contacts over time  Transient Connectivity – formation of transient connected subnets (TCS) for periods of time  Transient Community Structure – different communities created through the day  Show that these perspectives have predictable behavior representable by Gaussian functions  Develop forwarding metrics based on these functions to use in a forwarding strategy for better data delivery

Overview: Forwarding Algorithm  Forwarding decision of whether node i sends data, d k, to node j dependent on node forwarding metrics and forwarding strategy  m j = data forwarding metric of node j  Q i = strategy based metric of i to compare with m j  Common strategy that forwards data d k to node j if:  Node j is the destination node  Else if node j is in the community of the destination node and node i is not  Else if Q i < m j  Calculate m based on transient contact perspectives

Overview: Transient Perspectives  These perspectives provide more accurate estimation on the node's capability of contacting others within a given scope and time period  Fig (a) shows that λ, rate of contacts, varies over time and transient values provide greater fidelity than cumulative  Fig (b) shows how B may be a better choice than C due to indirect access to more nodes despite a lower contact rate

Overview: Transient Perspective  Rates are further refined by considering scope over time  Rate is weighted higher when node is in a community local to the destination node  For example, if the transient community structure of C is not considered, then λ t of node C would be 2.83 ((2x1+3x5)/6) and C would look better than A

9 Trace-Based Pattern Formulation  Performed study on multiple networks to understand and characterize their transient contact patterns  Network model and assumptions include  Contacts are symmetric  Stochastic contact process modeled as edge  Data is small such that bandwidth and buffering are considered irrelevant -Bluetooth devices detect peers nearby and make contact to them -WiFi search access points (AP) and make contact with others on same AP

10 Pattern Formulation: Transient Contact Distribution On-period of length L on is when there are a set of contacts within a threshold time T on Stable and predictable on- periods

11 Pattern Formulation: Transient Contact Distribution For T on set to 8 hrs, results are:

12 Pattern Formulation: Transient Contact Distribution  Graphs show that the distribution of on and off periods can be accurately approximated by normal distribution using mean and variance below  Model validated by mean on and off adding to 24 hrs, and >80% of contacts occur during on-periods

13 Pattern Formulation: Transient Connectivity  A node's connectivity is represented by the size of it's TCS (Transient Connected Subnet)  The TCS of node i during time period [t1,t2] consists of all nodes that have end to end comms with node i during that period 

14 Pattern Formulation: Transient Connectivity  TC depends on distribution of contact duration.  MIT: 20% > 1hr, UCSD: 30% > 1hr, Infocom: 5% > 30mins.  The average TCS size of each node.  MIT: over 50% > 3, UCSD: over 50% > 100, Infocom: negligible due to 30min issue above. 

Pattern Formulation: Transient Connectivity  The average TCS of all nodes can be approximated by :  Fig 3 & Fig 9 correlate therefore demonstrate that TC is proportional to the amount of contacts during time period t * A = amplitude function

Pattern Formulation: Transient Community Structure  Community structure only exists if there are more nodes than a certain threshold that form a stable community  Community relationship defined as a “joint-period” when a pair of nodes are in the same community  Detection of communities by k-clique and modularity method  Fig. 10 shows low community change at peak of node contacts (see Fig. 3 ) when community is stable and at night when only few contacts occur

Pattern Formulation: Transient Community Structure 17  Joint period can be also accurately approximated by normal distribution  Note μ co is less than μ on  Low variance indicates large communities

Forwarding Metric  Data forwarding metric based on node centrality  Measure node centrality for a given scope and time constraint using transient contact distribution and transient connectivity  C i is node i's centrality calculated by the sum of c ij, the number of nodes i can contact by contacting j  Direct contacts determined from transient contact distribution  Indirect contacts through j based on transient connectivity of node j

Forwarding Metric:Incorporating Transient Contact Pattern  For each pair of nodes i and j, the parameters of their on-period and off-period are updated every time they directly contact each other  Each node detects its TCS when contacted by broadcasting a detecting beacon  Transient connectivity is then updated by Gaussian curve fitting based on the recorded TCS sized during different hours

Forwarding Metric: Update Algorithm 20

Forwarding Metric: Contact Probability 21 The contact process is stable and predictable only during on-periods as in case 1(a) and 2 (b) Contact occurrence probability p ij =p ij 1 +p ij 2 Probability of contact during on-period p c (t 1,t 2 )=1-e -λ(t2-t1)

Forwarding Metric: Contact Probability Case 1 Case 2

Forwarding Metric: Incorporating Transient Connectivity Case 1 Case 2 Incorporate TCS where size given by: p ij from last page but now: Similar transformation finally,

Forwarding Metric:Prediction Error 24  Contact occurs but not known to be the start of an on-period or still an off-period  But 80% of contacts occur during on-period according to previous results  Long off periods lower accuracy of Case 2

25 Exploiting Community Structure  As in case (b) above, the forwarding metric is weighted by community membership over time  Network periodically detects community members  Can incorporate joint-period statistics

Performance Comparison: Setup  Used the test networks and randomly picked source and destination nodes  Social contact patterns are characterized real time as described  Community structure measured by modularity method  Performance criteria are data delivery ratio and forwarding cost

Performance Comparison: Setup  Compared with other forwarding metrics:  Contact counts (CC)-calculated cumulatively since network start  Betweenness-social importance of a node facilitating communication among others  Cumulative contact probability (CCP)-prob of contacting others based on cumulative contact rates  Forwarding Strategies used  Compare-and-forward-forward to all nodes with higher metric than itself  Delegation forwarding-forward to all nodes with higher metric than the highest it has ever had  Spray-forward limited set of copies to nodes with highest metric, each relay node forwards one copy to highest  Epidemic is the benchmark and BUBBLE Rap also tested

Performance Comparisons – Data Delivery Ratio 28  Using compare and forward strategy with different metrics  When time constraint is short, transient approach far outperforms all others and matches epidemic  With a longer time constraint, the cumulative characteristics become more consistent and transient advantage decreases

Performance Comparisons – Forwarding Costs  Graphs show transient metric results in 20% lower forwarding cost  General uptrend with increasing time constraint since more time allows more to be forwarded nodes to

Performance: Impact of T on 30  T on of 8 hrs has optimal performance  Smaller time contraint is more sensitive to sub-optimal T on

31 Performance: Impact of Transient Connectivity  Transient metric C i is calculated considering direct contacts only  Performance still better at lower time constraint, worsens at higher time  Delta performance across networks due to large TCS size in USCD and short contact duration in Infocom

32 Performance: Case 1&2 Contact Prediction  Case 1 predicts an on-period will continue and contributes more to delivery with low time constraint  Case 2 predicts future on-periods and has more accuracy with a longer time constraint

33 Performance: Static Community Structure  Using a static community structure shows decreased delivery with a high time constraint  With low time constraint, most delivery must be local anyway

Performance: Community detection Comparison of community detection methods show little difference in cost but better performance with modularity

Performance: Transient Community Structure with Different Forwarding Strategies  Delegation best overall with near max performance and lower cost since it's forwarding is more selective  Spray has limited cost and limited performance due to limited node

Conclusion  Transient social contact patterns are an effective way to determine a forwarding metric  Demonstrated predictive behavior of social contact patterns  Developed transient forwarding metric based on transient social contact pattern parameters  Evaluated forwarding performance and showed improved performance over static methods