1 / 31

Broadcast Scheduling on Mobile Ad Hoc Network

Broadcast Scheduling on Mobile Ad Hoc Network. Group 4 Yilin Shen , Zheng Li, Bharat C., Yan Qiao. Problem. An accident on a highway Evacuation is needed ASAP. A vehicle who received evacuation message will broadcast it to others. Main concern: latency

paley
Download Presentation

Broadcast Scheduling on Mobile Ad Hoc Network

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Broadcast Scheduling on Mobile Ad Hoc Network Group 4 YilinShen, Zheng Li, Bharat C., Yan Qiao

  2. Problem • An accident on a highway • Evacuation is needed ASAP. • A vehicle who received evacuation message will broadcast it to others. • Main concern: latency • A rescue team after an earthquake. • No fixed APs. • Rescuers must communicate to all team members. • Days of searching maybe possible. • Main concern: Energy efficiency.

  3. Advantage & Challenges • Advantage • Wireless Broadcast Advantage • Challenges exist in MANAT. • Interference • Synchronization • Overhead reduced by mobility • Why existing work isn’t satisfying? • Come later….

  4. Models (informally) • Network model • Each node has a transmission range r and a interference range f, r<f. • Nodes can have varied r and f. r v1 f

  5. Models (informally) contd. • Interference model • Mobility model • RPGM, RWK, RWP… u u v1 v2 v1

  6. Objectives • Minimize latency • Latency: time delay for the message to propagate over the network. • Minimize consumed energy • Can be calculated from the total time for transmitting signals.

  7. Approach1: Global Info Available • Assumption: • 1. global information is available before each broadcast. • 2. Network topology won’t change during a single broadcast. • Method: Hexagon tiling && distance-d hexagon coloring && dominating set.

  8. An example Naïve solution E E E E E B B B B B F F F F F A A A A A C C C C C G G G G G D D D D D Our ideal solution

  9. How to schedule? • Message propagates from layer to layer. • Layer: BFS, k-hop neighbors form the k-th layer • Schedule the transmission inside and between each layer. How? • 1. Dominating sets: • First transmit to dominating set (DS). Then DS to others. • 2. Coloring the nodes in DS: • If two nodes interfere, give them different colors. • Only nodes with the same color can transmit together. • Next question: how to decide the colors?

  10. Back to the example E E E E E layer 1 B B B B B layer 2 F F F F F A A A A A C C C C C G G G G G D D D D D

  11. How to do coloring? • In a simple graph, 5 colors are enough. Does it work here? • No. • Since nodes have different transmission range and interference range, we must be more careful. C B A

  12. Hexagon tiling & coloring • One way out – titling the space and color it. • (We borrow the concept from cellular network)

  13. Shorting comings of approach 1 • Maintaining the global information in MANAT is even costly than broadcast itself. • Even if only a small portion of nodes keep such global information, the overhead is still high. • However, the BFT tree structure in this approach is the basis for following approaches.

  14. Approach 2: localized algorithm • Assumption: • 1. Each node knows its own location. • 2. Network topology won’t change during a single broadcast. • Method: localized tiling and coloring method.

  15. The Localized Algorithm • Each node locally identifies the hexagon it’s located in along with its color. • In each cellular (hexagon), select a master node to receive the message from previous layer and propagate it inside the hexagon. • All the master nodes form an overlay network.

  16. Update of local information • If a new node steps into hexagon: send a JOIN msg to the master node. • If a node (not master) leaves hexagon : send a LEAVE message to master node. • If a master node leaves hexagon: send a MSTRLEAVE msg to all nodes in the hexagon, select a new master. • Selecting a new master: different strategies: e.g. slowest speed / strongest power / central location…

  17. Approach3: Adaptive Topology Change • Condition: Nodes move faster than the transmission (connection-oriented) • In large-scale WSN/DTN, reliable transmission is more costly than in wired network • Retransmission will be needed more times and also take more time • Topology changes on different periods of timestamps. How to HANDLE this scenario?

  18. Adaptive Topology Change • There can be one coordinator / server to do the broadcast scheduling with the global information. • Problem if using the previous approach: • Q: The BFS tree is still the same? • A: No, some edges may not exist any more because of the transmission range changing.

  19. Adaptive Topology Change (Cont.) • Q: How to modify the previous approach? • Q1: Is the coloring information still valid? • A1: Yes. Given a 2D space and a tiled hexagonal lattice, the coloring information is still valid with respect to hexagon.

  20. Adaptive Topology Change (Cont.) • Q2: How can the mobile node know the coloring information? • A2: Intuitively, we can equip each node with a GPS such that each node will know each own location. For the specific location information and hexagonal lattice coloring, each node will know each own color.

  21. Adaptive Topology Change (Cont.) • Q3: If the nodes do not have GPS, what will happen? • A3: With the knowledge of mobile models we discussed in class, we have get the location information based on its original location, velocity, destination (RWP), or based on its leader’s information(Leader will have GPS) (RPGM).

  22. Adaptive Topology Change (Cont.) • Q4: The interference can be avoided? • A4: Yes. Since the transmission is based on the coloring information. Since the coloring keeps the same for some specific hexagonal lattice, the broadcast will be interference-free.

  23. Adaptive Topology Change (Cont.) • Q5: With the change of the topology, the BFS does not hold any more. How can we handle it? • A5: We will construct a forest by BFS. • Notice that we will have a benefit here. From the original phase, the reason of the BFS tree is because only one node has the message.

  24. Adaptive Topology Change (Cont.) • Now there are more than one node have the information. In this case, we can construct a forest (a number of trees) instead of only one tree.

  25. Adaptive Topology Change (Cont.) • Q7: What is the benefit with more than one BFS? • A7: There will be several points of view: • For a smaller tree with lower breadth, the topology will be most likely to keep the same, we can do approach 1 directly. • All trees may be able to do the broadcast simultaneously without interference. (TO DO)

  26. Adaptive Topology Change (Cont.) • Q8: How can we construct the forest? • A8: We may use the RFP (Reverse Forwarding Path) method for the construction. Each node can have one bit to indicate whether it has already belonged to one tree or not. • TO DO: Is there a more intelligent construction such that all trees can broadcast simultaneously?

  27. Synchronization • Q9: Can the global clock also be used for synchronization? • A9: We consider three cases (from ideal to practical): • If the behaviors of all nodes are the same, i.e. moving velocity -> global clock • If the behaviors of all nodes are very similar -> global clock can be used with a little adjustment

  28. Synchronization(Cont.) • If neither of the above, we can use Lamport’s algorithm for the synchronization • (Notice that here we just need to do the synchronization for each tree rather than for all nodes)

  29. Comparison between Models • RWP • Advantage: • Easy to know location with velocity and destination • Disadvantage: • Difficult for synchronization • Difficult for forest construction (not possible to be adaptive) • RPGM • Advantage: • Easy to construct forest • Easy for synchronization since a group of nodes move similarly • Disadvantage: • Moving limitation

  30. Related works • Next speaker: Zheng Li

  31. Thank you!

More Related