A Distributed Weighted Cluster Based Routing Protocol in MANETs International Conference on Information Telecommunication & Computing (ITC 2011) Presented by:- Naveen Chauhan
To study various cluster based routing schemes in mobile ad-hoc networks and implement distributed weighted cluster based routing algorithm. Objective
A lot of research is currently going on in mobile ad-hoc networks
chief focus being to develop an efficient routing protocol which provides for efficient communication with minimum energy requirement
Among these schemes, the cluster based routing schemes are of particular interest as they lead to saving of energy and communication bandwidth leading to more efficient routing.
Some of these cluster based routing schemes are CBRP, weighted clustering algorithm (WCA), distributed weighted clustering algorithms etc.
Routing is a very challenging task in mobile ad hoc networks due to their peculiar characteristics. Among various routing schemes Cluster based routing schemes are of the most interest as they give better performance while conserving energy.
Cluster based routing schemes work by forming clusters of nodes with each cluster having its own clusterhead with the clusterhead being responsible for routing. Nodes are grouped into clusters to reduce communication overhead. Intra cluster communication takes place inside each cluster via direct links and does not involve nodes from other clusters even if they are closer. All inter-cluster communication relays through the cluster head.
Cluster based routing schemes work by forming clusters of nodes with each cluster having its own clusterhead with the clusterhead being responsible for routing.
Nodes are grouped into clusters to reduce communication overhead.
All inter-cluster communication relays through the cluster head.
Thus a virtual network infrastructure is created which resembles fixed network infrastructure.
Before Clustering After Clustering
Mobile Ad-Hoc Networks (MANETs)
MANETs or Mobile ad-hoc networks are a form of wireless networks which do not require a base station for providing network connectivity
The idea is to form a totally improvised network that does not require any pre-established infrastructure
Each node acts as a host and a router at the same time
MANET relies on the cooperation between its participating members
Figure 2.1 A Simple Mobile Ad-hoc network
Routing Protocols in MANETs
Destination-Sequenced Distance Vector routing (DSDV) is a table driven routing protocol. In DSDV, each mobile node in the network maintains a routing table with entries for every possible destination node, and the number of hops to reach them. The routing table is periodically updated for every change in the network to maintain consistency.
Dynamic source routing (DSR) protocol is an on-demand routing protocol and it maintains a route cache, which leads to memory overhead. DSR has a higher overhead as each packet carries the complete route, and does not support multicast.
Ad Hoc On-Demand Vector Routing (AODV) protocol is a reactive routing protocol for ad hoc and mobile networks that maintain routes only between nodes which need to communicate. AODV is an improvement on DSDV because it typically minimizes the number of required broadcasts by creating routes on an on-demand basis, as opposed to maintaining a complete list of routes as in the DSDV algorithm.
Cluster Based Routing
In cluster-based routing, a virtual network infrastructure must be created through the clustering of nodes in order to provide scalability
Each cluster can have a cluster head, which is responsible for intra- and inter-cluster coordination in the network management functions
Nodes inside a cluster communicate via direct links
Inter-cluster communication is performed via the cluster-heads
The stable clustering of nodes is the key to create this infrastructure
Cluster Based Routing Schemes for MANETs
Highest Degree Heuristic:
In this scheme degree for each node is computed based on its distance from others. The node with maximum number of neighbours is chosen as the cluster head
In this scheme random ids is generated and uses a heuristic function to assigns a unique id to each node and choosing the node with the minimum id as a cluster-head.
In this each node is assigned weights based on its suitability of being a cluster-head. A node is chosen to be a cluster-head if its weight is higher than any of its neighbour’s weight; otherwise it joins a neighbouring cluster-head.
Weighted Clustering Algorithm (WCA)
In this algorithm there are several system parameters like the ideal node degree, transmission power, mobility and the battery power of the nodes.
Depending on specific applications, any or all of these parameters can be used in the metric to elect the cluster heads.
The final weight of the node is determined by taking the combined weight of all individual parameters.
The node with the lowest weight becomes clusterhead.
Parameters used are degree difference, sum of distances, mobility, time spent as clusterhead etc.
Distributed Clustering Algorithm (DCA)
The Distributed Clustering Algorithm is a modified version of the Lowest-Identifier algorithm.
For each cluster, it chooses a node with locally lowest ID among all the neighbouring nodes as a cluster head.
Every node can determine its cluster and only one cluster, and transmits only one message during the algorithm.
A node decides which role to assume only when all its neighbors with bigger weights have decided their own roles.
Distributed Weighted Clustering Algorithm(DWCA)
This algorithm is an enhanced version of WCA to achieve distributed clustering set up and to extend lifetime span of the system.
In addition to distributed algorithm, weight is computed like in WCA for each node instead of taking random value as in DCA.
DWCA consists of the clustering set up and clustering maintenance phases.
It has better battery management because consumed battery power is taken as a parameter in selection of clusterhead.
Distributed Score Based clustering Algorithm (DSBCA)
Similar to DWCA, but instead of weight score is kept for each node.
The score of node N is defined as Equation.
Score= ((Br×C1) + (Nn×C2) +(S×C3) + (Nm×C4))
Here C1, C2.. are score factors for following parameters:
Battery Remaining, Br
Number of neighbours, Nn
Numbers of members, Nm
Stability, S : the time spent by neighbours of node in its vicinity
Random number of mobile nodes with each node having fixed energy and random mobility.
The transmission range of each node can be specified and each node can pass messages to all the nodes in its transmission range.
Further each node starts with some energy and its energy decreases each time it passes a message. A node fails if its energy has all been consumed.
Mobility used is random mobility. A mobile node moves for a random interval in a random direction and then changes its direction randomly.
Each node calculates its own weight based on the following factors:
Node connectivity: The number of nodes that can communicate directly with the given node i.e. that are in its transmission range.
Battery Power: The power currently left in each node. The energy is consumed by sending and receiving of messages.
Mobility: Running average of speed of each node. If mobility is less, the node is more suitable to become clusterhead.
Distance: Sum of distance of the node from all its neighbours.
The node with maximum weight becomes clusterhead.
The weights associated with each parameters were determined by trial and error.
Weight Calculation Algorithm
1. Find connectivity c for each node which is the number of neighbours of each node.
2. Find the energy remaining, e for each node.
3. Compute the mobility M for each node which is the running average of the speed until the current time t.
4. Compute the sum of distances with all its neighbours, d for each node.
5. Calculate the combined weight W as
W= w 1 * c + w 2 *e – w 3 *M + w 4 *d
1. Each node finds its neighbours and builds its neighbourhood table.
2. Each node calculates its weight by calling the weight calculation algorithm given above.
3. Each node broadcasts its weight to its neighbours. If it has maximum weight among its neighbours, it sets the clhead variable to 1 otherwise, the clhead variable is set to 0.
4. The node with maximum weight broadcasts clhead message to other nodes.
5. On receiving a clhead message a node checks all the nodes from which it receives clhead message. The node with maximum weight becomes the clusterhead of that node.
6. Then we assign a random value between 0 and 1 to each node and a threshold is taken.
7. If the random value assigned to the node is greater than threshold value, then set mobile = 1, otherwise 0.
8. If for a node clhead=1,then set mobile =0.
9. If mobile =1, set value of direction randomly. Increment or decrement the value of x and y depending upon the direction to show mobility.
10. In case a new node is added, it calculates its weight by calling weight calculation algorithm and repeats steps 3, 4 and 5.
11. In case clusterhead fails, the algorithm is repeated.
IMPLEMENTATION AND RESULTS
Implemented in Omnet++
Results No of Nodes No of Clusterheads 1. . No of Clusterheads Formed vs No of Nodes
Figure depicts the average number of clusters formed with respect to the total number of nodes in the Ad-hoc network.
If the node density is increased, our algorithm produced constantly less clusters in comparison with the original algorithm.
As a result, our algorithm gave better performance in terms of the number of clusters when the node density and node mobility in the network are high.
Results 2. No of Control Messages vs No of Nodes No of Nodes No of Control Messages
Figure shows the overhead of packets generated per node during the initial clustering set up phase.
The overhead increased as the number of nodes increases.
Each node independently chose one of its neighbors with the highest score to be its cluster head and thus the cluster head selection was performed in a distributed manner with most recently gathered information of current state of the neighbors.
Results 3. No of Reaffiliations vs No of Nodes No of Nodes No of Reaffilations
This figure describes the number of reaffilations that are done when a node becomes mobile.
The nodes become mobile at a random value so this criteria is rather more for self evaluation.
The purpose of this factor is that the reaffilations must not exceed the factor which increases network overhead and fails the meaning of clustering process.
Results 4. Energy Left vs No of Nodes No of Nodes Energy Left
This graph shows the total energy remaining in the network. Initial energy is the total energy of all the nodes.
As the messages are passes the energy of a node decreases. Our protocol improves on the original protocol in the sense that it reduces the energy consumption.
As the number of nodes increases the number of messages increases and thus the energy consumed also increases.
CONCLUSION AND FUTURE WORK
In this project we have studied various cluster based routing schemes and implemented a novel distributed weighted clustering algorithm making some modifications and improvements on existing algorithms.
As demonstrated, our algorithm reduces the clusterhead formation and control messages overhead thus improving overall performance and reducing energy utilization.
Since energy utilization is the most important criteria in cluster based routing schemes, our protocol provides better results than existing distributed clustering algorithms.
For the next step in this protocol, new parameters can be added into weight computation of nodes so as to give even better performance.
Also, in this algorithm the clusterhead selection is limited to single hop neighbours. This protocol can be extended to include multi-hop or k-hop neighbours.
Also, security is an important issue which can be addressed in the future. Encryption and authentication etc. can be implemented in the protocol to make it more secure.
Since, this protocol has been tested on simulation software; it can be implemented in a real ad-hoc system to evaluate its performance in real world scenarios.
 Sahar Adabi, Sam Jabbehdari, Amirmasoud Rahmani, Sepideh Adabi, “A Novel Distributed Clustering Algorithm for Mobile Ad-hoc Networks”, Journal of Computer Science, pp 161-166, 2008.
 Wonchang Choi, Miae Woo, “A Distributed Weighted Clustering Algorithm for Mobile Ad Hoc Networks”, Proceedings of the Advanced International Conference on Telecommunications and International Conference on Internet and Web Applications and Services, 2006.
 Basagni, S., “Distributed clustering for ad hoc networks” In Proceedings of I-SPAN ’99, 1999, pp. 310-315.
 Chatterjee, M., Das, S., and Turgut, D., “WCA: a weighted clustering algorithm for mobile ad hoc networks,” Journal of Cluster Computing (Special Issue on Mobile Ad hoc Networks), 5, 2002, pp.193-204.
 Matthias R. Brust, Adrian Andronache and Steffen Rothkugel, “WACA: A Hierarchical Weighted Clustering Algorithm optimized for Mobile Hybrid Networks”, Proceedings of the Third International Conference on Wireless and Mobile Communications (ICWMC'07), 2007
 Mohamed Elhoucine Elhdhili, Lamia Ben Azzouz, and Farouk Kamoun, “Lowest Weight: Reactive Clustering Algorithm for Adhoc Networks”, IEEE Journal on Mobile Communications, 2006
 L. Ramachandran, M. Kapoor, A. Sarkar and A. Aggarwal, “Clustering algorithms for wireless ad hoc networks” , Proceedings of Workshop on Discrete Algorithms and Methods for Mobile Computing and Communications , Boston, August 2000, pp. 54–63.
 Y. Wang, H.R. Chen X.Y.Yang and D.Y. Zhang, “WACHM: Weight Based Adaptive Clustering for Large Scale Heterogeneous MANET,” IEEE Journal, 2008.
 M. Jiang, J. Li, and Y. C. Tay, “Cluster based routing protocol (cbrp),” draft-ietf-manet-cbrp-spec-01.txt,IETF, Internet draft version 01, July 1999.
 Xiaoguang Niu, Zhihua Tao, Gongyi Wu, Changcheng Huang, Li Cui, “Hybrid Cluster Routing: An Efficient Routing Protocol for Mobile Ad Hoc Networks”, Communications, IEEE International Conference, vol. 8, pp 3554-559 (2006)
C. R. Lin and M. Gerla, “Adaptive Clustering for MobileWireless Networks,” IEEE J. Select. Areas Commun.,vol. 15, pp. 1265 – 1275, Sept. 1997.
 F. Baker “ An outsider's view of MANET draft-baker manet review” , Network Working Group March 17, 2002
 C.Perkins,S.Das, “Ad hoc On-Demand Distance Vector (AODV) Routing” , Network Working Group July 2003
 Sanjay Kumar Dhurandherl and G. V. Singh, “Power Aware Clustering Technique in Wireless Ad Hoc Networks”, IEEE Journal on Mobile Communications, 2006