1. TOPOLOGY DISCOVERY IN SENSOR NETWORKS Budhaditya Deb, Sudeept Bhatnagar Badri Nath Department of Computer Science, Rutgers University, May 2001.

2. WHAT ARE SENSOR NETWORKS ? Distributed and highly complex network systems Distributed and highly complex network systems Hundreds of tiny sensor nodes Hundreds of tiny sensor nodes Behavior is highly unpredictable Behavior is highly unpredictable Randomness in individual node state and network structure. Randomness in individual node state and network structure. Used for automated information gathering. Used for automated information gathering.

3. CONSTRAINTS FACED BY NODES IN SENSOR NETWORKS Sensor network nodes are resource constrained Sensor network nodes are resource constrained Energy is limited by battery power Energy is limited by battery power Form factor limits the amount of resources that can be put in a node Form factor limits the amount of resources that can be put in a node Protocols designed for sensor networks need to be highly efficient and optimized in terms of resources they consume. Protocols designed for sensor networks need to be highly efficient and optimized in terms of resources they consume.

4. The TopDisc Algorithm(Overview) Because of the above reasons, regular SNMP functionalities cannot be used over wireless sensor networks Because of the above reasons, regular SNMP functionalities cannot be used over wireless sensor networks Paper describes a TopDisc algorithm which is a topology discovery algorithm used for Paper describes a TopDisc algorithm which is a topology discovery algorithm used for 1. Data dissemination 2. Data aggregation 3. Duty cycle assignments 4. Network State Retrieval

5. Overview of TopDisc A set of distinguished nodes is obtained. A set of distinguished nodes is obtained. These nodes only reply to the topology discovery probes. These nodes only reply to the topology discovery probes. These nodes form clusters comprised of the nodes in their neighborhood. These nodes form clusters comprised of the nodes in their neighborhood. These clusters form a tree structure called TreC, rooted at the monitoring node. These clusters form a tree structure called TreC, rooted at the monitoring node.

6. Models for sensor networks Models depict the current state of the network. Some useful models would be : Network Topology Network Topology Energy Map Energy Map Usage Pattern Usage Pattern Non-deterministic Models Non-deterministic Models

7. Network Management Functions The models described above could be used for many network management functions like : The models described above could be used for many network management functions like : Deployment of sensors Deployment of sensors Setting Network Operating Parameters Setting Network Operating Parameters Monitor Network States using Network Models Monitor Network States using Network Models Network Maintenance Predict Future Network States Design of Sensor Networks

8. TopDisc Algorithm The TopDisc Algorithm can be divided into three stages of execution : A monitoring node requiring the topology of the network initiates a "topology discovery request". A monitoring node requiring the topology of the network initiates a "topology discovery request". This request diverges throughout the network reaching all active nodes. This request diverges throughout the network reaching all active nodes. A response action is set up which converges back to the initiating node with the topology information. A response action is set up which converges back to the initiating node with the topology information.

9. Topology Discovery Three methods can be used for topology discovery Three methods can be used for topology discovery Direct Response Direct Response Aggregated Response Aggregated Response Clustered Response Clustered Response

10. TopDisc Algorithm Two different approaches are used to select cluster heads : Three color approach Three color approach Four color approach Four color approach

11. The three color approach The three different colors used and what they mean White: Yet undiscovered node White: Yet undiscovered node Black: Cluster head node Black: Cluster head node Grey: Node which is covered by black node Grey: Node which is covered by black node

12. The Three Color Approach Initiate topology discovery  Black Initiate topology discovery  Black ReceiverSender Converts to Action Taken WhiteBlackGreyBroadcast WhiteGrey Black(With random del) From black, becomes grey Black, Grey XXXX-------------- Ignore requests

13. The Three Color Approach(contd)

14. The Four Color Approach The four different colors used and what they mean : White: Yet undiscovered node. White: Yet undiscovered node. Black: Cluster head node Black: Cluster head node Grey: neighbor of a black node. Grey: neighbor of a black node. Dark Grey: Discovered node, which currently is not covered by any neighboring black node. Dark Grey: Discovered node, which currently is not covered by any neighboring black node.

15. The Four Color Approach (Contd) Initiate topology discovery  Black Initiate topology discovery  Black Grey or black  IGNORE Requests Grey or black  IGNORE Requests

16. Four color approach ReceiverSender Converts to Action taken WhiteBlackGrey Broadcast with delay WhiteGrey Dark Grey Broadcast, Start timer White Dark Grey Black With random delay, if packet from black--grey

17. Four color (Example)

18. TopDisc Response Mechanism Each node has the following information at the end of the topology discovery period: Cluster identification  Black node Cluster identification  Black node Grey node knows its cluster id. Grey node knows its cluster id. Each node knows its parent black node. Each node knows its parent black node. Each black node knows the default node to which it should forward packets to reach the parent black node. Each black node knows the default node to which it should forward packets to reach the parent black node. All nodes have their neighborhood information. All nodes have their neighborhood information.

19. TopDisc Response Using the above information, the steps for TopDisc Response are described as follows. When a node becomes black  starts timer to receive information from its children When a node becomes black  starts timer to receive information from its children Aggregates information  forwards to its parent Aggregates information  forwards to its parent All forwarding nodes in between black nodes may also add their adjacency lists to the list from black nodes. All forwarding nodes in between black nodes may also add their adjacency lists to the list from black nodes.

20. Applications 1. Retrieving Network State Connectivity Map Connectivity Map Reachability Map Reachability Map Energy Model Energy Model Usage Model Usage Model 2. Data Dissemination and Aggregation 3. Duty Cycle Assignment

21. Applications(contd) 1. Assignment with location information information 2. Assignment without location information information

22. Conclusion Further work and research on TopDisc to come up with sNMP (Sensor Network Management Protocol). Further work and research on TopDisc to come up with sNMP (Sensor Network Management Protocol).