Presentation is loading. Please wait.

Presentation is loading. Please wait.

A Topology Discovery Algorithm for Sensor Networks Go Suzuki CS691, SSNS Spring 2003.

Similar presentations


Presentation on theme: "A Topology Discovery Algorithm for Sensor Networks Go Suzuki CS691, SSNS Spring 2003."— Presentation transcript:

1 A Topology Discovery Algorithm for Sensor Networks Go Suzuki CS691, SSNS Spring 2003

2 Introduction Background for Wireless Sensor Network TOPLOGY DISCOVERY Algorithm Simulation & Results Conclusions

3 Wireless Sensor Network Background MEMS Technology Low Power MAC Protocols Power Aware Routing Algorithm Energy Efficient Communication Protocol Tiny Diffusion

4 Sensor Network Management possible models : Network Topology connectivity / reachability map Energy Map energy levels of the nodes at different path Usage Pattern network activity, data transmit / unit time Cost Model equipment / energy / human costs Non-deterministic Models statistical & probabilistic models

5 Topology Discovery? Goal: To construct the topology of the whole network from the perspective of a single node. Reducing the communication overhead of the process.

6 Topology Discovery Algorithm: 3 stages Monitoring node  Send “topology discover request” Divergence of requests  reaching all active nodes Response action  topology information back to the initiating node

7 Topology Discovery Overview of TopDisc Response 1. Direct Response Node B replies back to node A Node C replies to node B; node B forwards the reply to node A Node D replies to node B; node B forwards the reply to node A Node A gets the complete topology ! A B CD

8 Topology Discovery Overview of TopDisc Response 2. Aggregated Response Node C and D forward request; node B listens to these and deduces them to be its children Node C replies back to node B; node D replies back to node B Node B aggregates information from C, D and itself; node B forwards the reply to node A Node A gets the complete topology ! A B CD A B CD Listen Agg. RES

9 Topology Discovery Overview of TopDisc Response 3. Clustered Response Assume: node B is a cluster head and nodes C and D are part of its cluster. Node C and D do not replay Only node B replies to node A Node A does not get link C  D A B CD A B CD A B CD RES Cluster head B

10 Cluster Response Approaches Cluster head 1. V =  V i 2.  x  V i, edge( x, i )  E graph: G(Vertex, Edge) Vi be the neighborhood list of node i, with i  C

11 White: undiscovered / not receive packet Black: Cluster head node Grey: node which is covered at least once black node Cluster Response Approaches Request Propagation with 3 colors

12 Coloring algorithm 1. Using coloring mechanism to find the required set nodes. 2. Using a forwarding delay inversely proportional to the distance between receiving and sending node.

13 a: initial state a: broadcast request to b & c b farther to a  wait shorter c closer to a  wait longer ( forwarding delay ) b: broadcast request to e & c e closer to b  wait longer c farther to b  wait shorter ( forwarding delay ) Expand range as soon as possible ( depends on density ) Cluster Response Approaches Request Propagation with 3 colors

14 White / Black / Grey: same condition as before Dark Grey: Discovered node,  which currently is not covered by any neighboring black node and hence is two hops away from a black node.  White node changes to dark grey on receiving a request from grey.  Timer to become black node get req from black  grey expired w/o req  black Cluster Response Approaches Request Propagation with 4 colors

15 a: initial state a: TopDisc request to b b: TopDisc request to c & e c farther to b  wait shorter w/ dark.G e closer to b  wait longer w/ dark.G ( forwarding delay ) ( timer starts to become black ) c: TopDisc request to d Expand range as soon as possible (depends on density) Cluster Response Approaches Request Propagation with 4 colors

16 Advantage: # of clusters is less than with 3 colors clusters are formed with lesser overlap solitary black nodes (time out D.grey nodes with no neighbors) though number of black nodes is similar to three-color case, the number of bytes transmitted is lower. Cluster Response Approaches Request Propagation with 4 colors

17 Cluster Response Approaches TopDisc Response Mechanism(TreC) 1. Node becomes black  sets up a timer  wait for the discovery request from children black nodes

18 2. Forwards aggregates all neighborhood lists  all neighborhood list from its children/itselt  when timer for ACK expires, forward aggregated neighborhood list Cluster Response Approaches TopDisc Response Mechanism(TreC)

19 3. All forwarding nodes in between black nodes may also add their adjacency list to the list from black nodes Cluster Response Approaches TopDisc Response Mechanism(TreC)

20 Tree-Cluster(TreC) for 200 nodes ***Timeouts of ACK should be properly set. Timeouts of children black nodes should always expire before a parent black node.

21 Cluster Response Approaches Information of each nodes  Clusters is identified by the black node  A grey node knows its cluster ID  Each black node knows the default node  All nodes have their neighborhood information

22 Cluster Response Approaches Handling Channel Errors TopDisc request: would not be a problem because of flooding packets (packet Losses  # of black node increase ) Topology ACK: serious problem because of single path to return back to sink.  assume links are symmetrical ( nodes listen neighbors transmit )  Packet has to be stored at a node till the packet is reliably transmitted Indirect ACK mechanism for reliable transmission. BA Liste n ? ?

23 Cluster Response Approaches Characteristics of clusters The total surface area and the communication range of nodes bound the maximum number of black nodes formed. Number of nodes in each cluster depends on the local density of network Depth of tree is bounded Routing paths are near optimal for data flow between sour and sink.

24 Applications of TopDisc Retrieving Network State Connectivity Map Direct / Aggregate response  O.K., Clustered response method  x Reachability Map Connectivity map is a superset of the Reachable map Energy Model Each (black) nodes can cache energy info for all neighbors. Usage Model Cache receive / transmit rate and send its response

25 Applications of TopDisc Data Dissemination and Aggregation Each cluster has a minimal number of nodes.  active to transfer packets between a parent-child cluster pair The area covered propagate up the tree and the monitor covers the whole field

26 Applications of TopDisc Duty Cycle Assignment 1. Assignment with Location Information Cluster a (parent) a (black): TopDisc request to c c (grey) : TopDisc request to b b (black): will be child cluster p (mid-point) of parent/child node

27 Cluster a and Cluster b p sends a packet  a determines p is within range of c  otherwise c can listen to the packet from p. Node c forwards it to d *** since c is in range of p the black node a does not need to forward this packet. Applications of TopDisc Duty Cycle Assignment

28 Cluster a and Cluster b a sends packet  c get packet and forward it to nodes within its range b gets request with couple steps c: centermediate node between two black nodes 2. Assignment w/o Location Information Applications of TopDisc Duty Cycle Assignment

29 Simulations & Results Byte Overhead for TopDisc Byte Overhead ( Direct / Aggregate / TopDisc ) # of nodesCommunication range Only neighborhood list of Black nodes  stay low

30 Simulations & Results Average path length Path Length ( Shortest path / TopDisc )

31 Simulations & Results # of nodes sharing forwarding duty Number of node sharing forwarding duty Nearly 50% sharingNearly 40% sharing # of black nodes + # of default nodes

32 Conclusions TopDisc selects a set of distinguished node TopDIsc constructs a reachability map TopDisc logically organizes the network and forms TreC TreC for -- efficient data dissemination & aggregation -- duty cycle assignment -- network state retrieval Completely distributed, used only local information and is highly scalable

33 References A Topology discovery Algorithm for sensor Networks with Applications to Network Management Benjie Chen, Kyle Jamieson, Hari Balakrishnan, and Robert Morris, Span: An Energy-Efficient Coordination Algorithm for Topology Maintenance in Ad Hoc Wireless Networks, ACM/IEEE International Conference on Mobile Computing and Networking (MobiCom 2001), Rome, Italy, July 16-21, 2001. Alberto Cerpa and Deborah Estrin, ASCENT: Adaptive Self-Configuring Sensor Networks Topologies, International Annual Joint Conference of the IEEE Computer and Communications Societies (INFOCOM 2002), New York, NY, USA, June 23-27 2002.

34 Questions / Comments?


Download ppt "A Topology Discovery Algorithm for Sensor Networks Go Suzuki CS691, SSNS Spring 2003."

Similar presentations


Ads by Google