1. W IRELESS A D H OC N ETWORKS Vamsi Paruchuri Associate Professor Department of Computer Science University of Central Arkansas

2. Outline Wireless networks – Brief History and Future – Some Current Technologies Cellular, WLAN, Bluetooth, AdHoc, Sensor Broadcasting in Wireless AdHoc Networks – Challenges – Geometric Approach – 3-Dimensional Networks – Summary Vamsi Paruchuri University of Central Arkansas 2

3. Some history Ancient Systems: Smoke Signals, Carrier Pigeons, … Radio invented in the 1880s by Marconi Many sophisticated military radio systems were developed during and after WW2 Cellular has enjoyed exponential growth since 1988 – Almost 1 billion users worldwide today – Ignited the recent wireless revolution – Growth rate tapering off Vamsi Paruchuri University of Central Arkansas 3

4. Future Wireless Networks Vamsi Paruchuri University of Central Arkansas 4 Wireless Internet access Nth generation Cellular Wireless Ad Hoc Networks Sensor Networks Wireless Entertainment Smart Homes/Spaces Automated Highways All this and more… Ubiquitous Communication Among People and Devices Hard Delay Constraints Hard Energy Constraints

5. Cellular Systems: Reuse channels to maximize capacity Geographic region divided into cells Frequencies/timeslots/codes reused at spatially-separated locations. Co-channel interference between same color cells. Base stations/MTSOs coordinate handoff and control functions Shrinking cell size increases capacity, as well as networking burden

6. WLANs connect “local” computers (100m range) Breaks data into packets Channel access is shared (random access) Standards: 802.11a/b/g Vamsi Paruchuri University of Central Arkansas 6 Wireless Local Area Networks (WLANs) 01011011 Internet Access Point 0101 1011

7. Bluetooth Cable replacement RF technology (low cost) Short range (10m, extendable to 100m) 2.4 GHz band (crowded) 1 Data (700 Kbps) and 3 voice channels Widely supported by telecommunications, PC, and consumer electronics companies Few applications beyond cable replacement Vamsi Paruchuri University of Central Arkansas 7

8. Ad-Hoc Networks Peer-to-peer communications. No backbone infrastructure. Routing can be multihop. Topology is dynamic. Fully connected. Vamsi Paruchuri University of Central Arkansas 8

9. Sensor Networks Energy is the driving constraint – Nodes powered by non-rechargeable batteries – Data flows to centralized location. – Low per-node rates but up to 100,000 nodes. – Data highly correlated in time and space. – Nodes can cooperate in transmission, reception, compression, and signal processing. Vamsi Paruchuri University of Central Arkansas 9

10. Many Motivations for Wireless Unrestricted mobility Unplugged from power outlet Significantly lower cost No cable, low labor cost, low maintenance Ease Minimum infrastructure - scatter and play The bottom line is that the world of the future will have fairly ubiquitous wireless networking connectivity – how soon is the question that remains to be answered!

11. Ad Hoc Networks – Each mobile node operates not only as a host but also as a router. – No existing infrastructure – No centralized administration – Omni-directional transmissions: when a host sends a packet, all its neighbors will receive that packet. Vamsi Paruchuri University of Central Arkansas 11

12. Challenges Unattended ad-hoc deployment Very large scale Scarce energy and bandwidth resources High noise and fault rates Dynamic / uncertain environments High variation in application-specific requirements Vamsi Paruchuri University of Central Arkansas 12

13. More Challenges Hidden Node Problem: A and C cannot sense and thus are hidden from each other Vamsi Paruchuri University of Central Arkansas 13 B ? ABCD ? Exposed Node Problem: C cannot transmit to D when B is transmitting to A, even though the transmission from C to D may not interfere the transmission from B to A

14. Broadcasting Broadcast – Send a message to all nodes in the network – It is used in the route query process in several routing protocols – The broadcast is unreliable: no acknowledgement of any kind Blind Flooding – A straightforward approach for broadcasting is blind flooding Each node rebroadcast s the packet when it receives the packet for the first time. – Blind flooding will cause many redundant transmissions – broadcast storm problem redundant rebroadcasts contention problem collision problem Vamsi Paruchuri University of Central Arkansas 14

15. Design Goals Scalability: Ability to provide acceptable level of service even in the presence of a large number of nodes in the network Reliability: Ability to cope with packet losses and node mobility Energy efficiency: – limited power sources. – Minimize number of transmissions thus reducing energy consumption Simplicity Vamsi Paruchuri University of Central Arkansas 15

16. Related Work MCDS (minimum connected dominating set) – A subset of nodes is the dominating set : if every node in the network is either in the set or a neighbor of a node in the set. – NP-complete problem – Does not consider that rebroadcasts are unreliable Approximate MCDS – Similar to constructing Spanning Tree – Requires global network knowledge Vamsi Paruchuri University of Central Arkansas 16 Five-Queen Problem

17. Related Work: Probabilistic Scheme – Rebroadcast probabilistically A host always rebroadcasts with a certain probability P. When P = 1, this is flooding. A smaller P will reduce the storm effect. – This approach is also known as gossiping in distributed systems well analyzed – Probability of rebroadcast ≈ 0.4 to ensure reachability Vamsi Paruchuri University of Central Arkansas 17

18. Related Work (cont.) – Counter based schemes If a host has received a broadcast packet “> C” times, then do not rebroadcast. – Estimated Additional Coverage 1 time => 41% 2 times => 19% 3 times => 9% 4 times => 5% > 4 times, very little additional area Vamsi Paruchuri University of Central Arkansas 18

19. Related Work: Topology Aware Neighborhood information How to decide forwarding nodes – Dynamically, or Neighborhood base Scalable Broadcast Algorithm (SBA), Flooding with Self pruning – Statically, or Set cover base Multipoint relaying, Dominant pruning, Ad hoc Broadcast Protocol (AHBP) Vamsi Paruchuri University of Central Arkansas 19

20. Broadcasting: Challenges Unreliable – High bit error rate, collusions Mobile nodes – Topology is constantly changing Biggest challenge: Non-isotropic Range – Very few works mention this challenge Vamsi Paruchuri University of Central Arkansas 20

21. Broadcasting Our Solution: Geometric Approach – Simple, scalable, reliable and minimizes retransmissions – Vamsi Paruchuri, Arjan Durresi, "Broadcast Protocol for Energy- Constrained Networks," IEEE Transaction on Broadcasting. – Arjan Durresi, Vamsi Paruchuri, R. Kannan, S.S. Iyengar, "Optimized Broadcast Protocol for Sensor Networks," IEEE Transactions on Computers. Vamsi Paruchuri University of Central Arkansas 21

22. Background What is the minimum number of circles required to completely cover a given 2-dimensional space? – Kershner, R. The Number of Circles Covering a Set. Amer. J. Math. 61, 665-671, 1939 Vamsi Paruchuri University of Central Arkansas 22

23. Our Approach Vamsi Paruchuri University of Central Arkansas 23 S

24. Algorithm In real conditions, nodes need not be in special locations Our protocol tries to maximize the distance per hop In each hop Self- selection of best positioned nodes – A delay in transmission is introduced – The delay is proportional to the distance from the ideal positions – Therefore, nodes closer to ideal positions will transmit. The others will wait longer. If they hear a transmission, that means they do not need to transmit at all. Vamsi Paruchuri University of Central Arkansas 24 S1 2 3 11 111 21 6 5 4 112 31 41 51 61 Ideal Case

25. Some Results Vamsi Paruchuri University of Central Arkansas 25

26. More Results Vamsi Paruchuri University of Central Arkansas 26

27. More Results Vamsi Paruchuri University of Central Arkansas 27

28. 3D Communications Communications among airplanes. – Increase safety – High density of airplanes – Need to communicate among them Communication is tall buildings – Hospitals, offices, and so on. Most of the wireless communications protocols are optimized for 2D Vamsi Paruchuri, Arjan Durresi, Leonard Barolli, Makoto Takizawa, "Three Dimensional Broadcast Protocol for Wireless Networks," in International Conference on Parallel Processing (ICPP). Vamsi Paruchuri University of Central Arkansas 28

29. Broadcast Protocols Many other distributed broadcast protocols – Protocols that require two-hop neighbor knowledge: Source Based Algorithm (SBA), Dominant Pruning, Multipoint Relaying, Ad Hoc Broadcast Protocol, Lightweight and Efficient Network-Wide Broadcast Protocol, etc. The above protocols can be extended for 3D, but when the network is very dynamic (such as that of airplanes), the neighbor information becomes obsolete very quickly. Vamsi Paruchuri University of Central Arkansas 29

30. How to minimize the number of transmissions? Geometric problem: How to fill a given space with the minimum number of spheres? – Each transmission, in ideal conditions is a sphere We show that to completely cover the surface of a sphere S of radius R, FOUR spheres of radii R are needed such that the center of each covering sphere lies on or within the covered sphere The centers of the four spheres form a regular tetrahedron. Vamsi Paruchuri University of Central Arkansas 30

31. Four Spheres to fill a Sphere Using regular tetrahedra leads to a small fraction of the space not covered by the four spheres Such gaps are not a real problem for our protocol for the following reasons: – the gaps are very small; – in a real networks, 3DB adapts itself and does not follow strictly the geometric approach. – The geometric model is used only in ideal cases, therefore the gaps are irrelevant Vamsi Paruchuri University of Central Arkansas 31

32. 3DB We ask only specific nodes to transmit The Source S ask nodes at the center of the calculated four spheres to transmit In real conditions, at those four specific locations, most likely there is no node. Therefore, we would like that the nodes closer to those four specific locations to retransmit Nodes self select themselves Vamsi Paruchuri University of Central Arkansas 32

33. Simulation Results –Contention We considered flooding, protocols in which the broadcasting nodes proactively chose neighbors to rebroadcast (neighbor), and 3DB. Vamsi Paruchuri University of Central Arkansas 33

34. Simulation Results - Efficiency As shown, while the number of nodes is increased, in 3DB less nodes are required to transmit, because more nodes are closer to ideal positions Vamsi Paruchuri University of Central Arkansas 34

35. Simulation Results - Mobility The influence of mobility in performance is not relevant because the protocol uses the nodes which happens to be neighbor at that moment. Vamsi Paruchuri University of Central Arkansas 35

36. Broadcasting: Summary We propose geometric approaches for broadcasting – an adaptive protocol, also optimized for three dimensional wireless transmissions. – very scalable, lower number of transmissions compared to other solutions – reachability is very good – very simple and adapt itself to the existing conditions – performance remains very good in mobile networks and in presence of errors. – Most importantly – non-isotropic transmissions do not impede the performance Vamsi Paruchuri University of Central Arkansas 36