Impact of CBR Traffic on Energy Consumption in MANET

: Mobile Ad hoc networks (MANETs) are power constrained since nodes are operated with limited battery supply. If a battery of the node drains, its ability to forward the traffic gets affected results in reduced network lifetime. There has been considerable progress in the battery technology, but not in par with the semiconductor technology. There are various techniques adopt different approach to achieve energy efficiency. The proposed approach uses a cost metric for path selection, which is a function of residual battery and current traffic load at a node. The simulation has been carried out using QualNet simulator. The performance is based on average energy consumption for varying CBR applications.


INTRODUCTION
In Mobile Ad Hoc networks (MANET), routing is a major concern as the nodes are battery operated. Multipath routing approaches [8] are being introduced to overcome the limitation of single path routing. The paths chosen can be link disjoint or node disjoint. Node disjoint paths have no common nodes except source and destination. Link-disjoints have no common links but can have common nodes. Multipath approaches have several benefits such as higher utilization of bandwidth, lower end-to-end delay and higher network lifetime. It also provides load balancing by forwarding the traffic through multiple paths. An energy aware multipath routing protocol provides a tradeoff between energy consumption and other metrics, such as link reliability, network capacity, throughput and end-to-end delay. Many of the energy efficient techniques minimize the energy consumption by selecting energy efficient path. However, when some nodes on the path forward large amounts of traffic.

Related Work.
In Minimum Energy Routing (MER) authors [10] described that a node consumes the minimum amount of energy to get the packet to the destination with the knowledge of cost of the link.MER has higher routing overhead, but consumes minimum energy. AODV Multiple alternative Paths (AODV-MAP) [2] is another variant of AODV. It considers both fail-safe paths and disjoint path .The main idea of the protocol is to find more number of alternate paths. . Scalable Multipath On-demand Routing (SMORT) [3] is a multipath routing protocol based on AODV [5]. It minimizes the routing overheads by using fail-safe paths instead of node-disjoint and link-disjoint paths. AOMDV [4] is one of the extensions of AODV [7]. It searches the paths without loop and link-disjoint paths.

Energy Efficient Routing Protocol
The proposed scheme describes route cost metric, technique to minimize node over utilization and computation of transmission power.

Route Cost Metric:
The lifetime of a node is based on cost function includes Residual Battery (RB) power and Energy Consumption rate (EC). Let the energy consumption rate of a node u at time t is CRu (t) and its residual battery be RBu (t).Let LTu (t) be the lifetime of a node u at time t is given by Equation [1].
The energy consumption rate Cru (t) is given by Equation [2].
Where CRold and CRnew represents the last and newly calculated value of energy consumption rate respectively η(< 1) is a weight function.

Minimizing Node Overutilization:
A critical node may exhaust battery and die due to the heavy traffic. This affects and reduces the network lifetime. A node drops RREQ packets if the connection request between source-destination exceeds the limit and reset the connection limit to that destination to one for the critical node requirement. Thus the node will forward subsequent RREQ packets for the connection establishment.

Computation of Transmission Power:
A node calculates its transmission power for data packets based on next-hop node's location and mobility pattern. A node u compute the required transmission power to reach the next node v is given by Equation [3].
Where D is the Euclidean distance between u and v. ∆ is the expected variance of distance between u and v considering the mobility. β is the path loss exponent with 2 ≤ β ≤ 4 and C is a constant. Expected variance is given by Equation Where Current time is the time at which node u is computing its transmission power, Reply time is the time at which node u has received the RREP packet from node v, S N is the speed of node v.

Simulation Parameters and Results
We have evaluated the scenario with 60, 80 and 100 nodes for 30 CBR applications using Qualnet 4.5 simulator. The simulation parameters are as shown in the table 1 and the scenario in Fig [1]. The plots of average energy consumption for different node densities with varying CBR connections in transmit, receive and idle modes are as shown in figures [3] to [10].       It is observed that the average energy consumption is low when the number of CBR connections are less in transmit mode. The number of RREQ packets increases as the CBR is increased results in increased energy consumption at the receiver. In idle mode, all the nodes consume energy while sensing the channel. When the connection limit is increase, less number of nodes will be in idle mode and energy consumption decreases.

Conclusion
In this paper we have used route cost metric based on residual battery power for the path selection for the efficient routing which avoids node overutilization.As the CBR traffic is increased,battery gets drained results in reduced network lifetime.