1. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Vladimir Katz Alex Gorodetsky The (k, l) Coredian Tree for Ad-Hoc Networks Advisors: Michael Segal Amit Dvir

2. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 What is an Ad-Hoc Network? Distributed network of sensors, without a centralized mechanism. High scalability and mobility. Easy to deploy. All sensors propagate the information to a sink node.

3. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Ad Hoc Networks: Motivation for use. Military operation - for fast establishment of communication during the deployment of forces in unknown and hostile terrain. Rescue missions - for communication in areas without adequate wireless coverage. Law enforcement - for fast establishment of communication infrastructure during law enforcement operations.

4. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Multicast Routing: Clustering Hierarchical Routing example A Sensor Ad Hoc Network

5. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Multicast Routing: Clustering Finding a core node

6. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Multicast Routing: Clustering Cluster the area (k parameter)

7. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Multicast Routing: Clustering Building the sub-tree (l parameter)

8. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Multicast Routing: Clustering Backbone of the network

9. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Backbone Construction: Problem Definition Center TreeCore Tree Minimization property eccentricity – the maximal distance to/from the backbone to any other vertex. distancesum – The sum of all vertices’ distances to/from the backbone. Motivation for use Fast response from the facility.Minimum total communication cost to all nodes. It is in our interest to find the best tradeoff between the two, which will give the best result for both goals altogether.

10. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Backbone Construction: Presenting the Coredian Minimizes the centdian function A convex combination of –Core minimization property (distancesum) –Center minimization property (eccentricity) Constrains : k - Number of leaves in the backbone tree, l - The maximal diameter of the backbone tree. centdian(v) = λ·distsum(v) + (1- λ) ·eccent(v), 0≤λ≤1 λ = 1λ = 0 Center Core

11. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Project Objectives Study of the existing backbone building algorithms. Implementation of the algorithms, using the OMNET++ simulation environment. Statistical Data Collection and Performances Analysis.

12. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Simulation Example

13. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Simulation Example: k-Core Tree (k=4, l)-Core Tree

14. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Simulation Example: k-Center Tree (k=4, l)-Core Tree(k=4, l)-Center Tree

15. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Simulation Example: (k, l) -Coredian Tree (k, l)-Core Tree(k, l)-Center Tree (k, l)- Coredian Tree

16. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Statistical Data Gathering

17. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Further Research: Performance Comparison K-Mcore Minimizes the sum of distances to the backbone and backbone’s weight at the same time. Was expanded into k-MCoredian, similarly as k-Core was expanded into k-Coredian. The need to expand the idea of Coredian into considerations of backbone weight led us to compare its performance with the performance of another existing algorithm, k-MCore.

18. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Further Research: Maintenance Issues can affect only the last added path to the backbone tree, therefore we “discard” the last added path to the tree and perform one additional iteration of the “find endpoint” algorithm. Arrival of a new node to the network

19. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Further Research: Maintenance Issues A much more complicated issue. Still being analyzed, by trying to look at different cases of connections’ failure, and characterizing the changes needed to be done in each one of these cases. A connection falls down (weight = 0)

20. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008 Future Work Further development of scalability sensitive maintenance algorithm for (k, l) Coredian. Development of a backbone algorithm which efficiently copes with the mobility of Ad-Hoc network. Extending the Coredian idea into considerations of node placement, network lifetime criteria, and data traffic management.

21. The (k, l) Coredian Tree for Ad-Hoc Networks Department of Communication Systems Engineering, Ben-Gurion University of the Negev, 2008