Performance Of Communication Wsn Nodes And Source Computer Science Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Wireless Sensor Network (WSN) is gaining popularity to fill the gap between physical world data and virtual world data. WSN is being exploited with exciting applications in the areas of environmental monitoring, vehicular communication, emergency alarms and military purpose. It's predicted that in 5-10 years time, millions of sophisticated WSNs would exist in various parts of the earth [1]. Typically the sensors in WSN operate in ad-hoc mode which utilizes multi-hop routing algorithms.

This paper analyse the performance of communication between the WSN nodes and source. In ad hoc routing some packets has to travel via several nodes to reach the source. In such cases the packets have a higher possibility of packet loss and delay.

So to measure the performance between multiple nodes it's critical to analyse the throughput and packet loss ratio. In this research study scenarios has been simulated for a sensor network in NS2 and the performance has been compared by altering the number of hops in-between the source and destination. There have been quite a lot researches on altering the routing algorithm, increasing the distance between source and destination and then performing evaluation. But, t his paper focuses on performance evaluation by varying the number of nodes.

Typically WSNs are being used to record and report environmental measurers[2]. Micro-Electro-Mechanical Systems (MEMS) have offered low cost miniaturized sensors consuming less power to form large sensor networks for such purposes[3]. Those sensors are spread out on the field of interest to measure parameters such as humidity, temperature, density of light and so on. The sensors are densely deployed to avoid impacts of sensor node failures and to increase the quality of data capturing[2]. WSNs consist of widely deployed sensors called as source to capture data, process and synchronize the data with the sink. Sink is used to query data from sensors. Also, it could influence the behaviour of other nodes as well[4]. WSNs become really valuable on environments where other traditional network infrastructures cannot be implemented. However, these networks cannot exercise the traditional protocols in the design. They hold distinct characteristics and constraints which distinguish them from other traditional networks[2].

Primarily, WSNs topology is unique, highly variable and dynamic[4]. Generally, the design will be a star -tree topology with all the nodes connected via broadcasting. The central sink node is either connected in the form of hierarchical or flat to the nodes [4]. The topologies often vary due to mobility of nodes and node failure. Mostly the traffic flow will be upstream towards the central node from the sensors reporting to sink. Though, the sink node could also create traffic towards sensors for management purpose and to query for data[4]. Secondly, the communication will suffer from high latency, low quality, and limited bandwidth. Usually the message size will be small and will not have segmentation[3]. Also, the sensors don't have power supply when the batteries get depleted. So that the power hunger protocols and algorithms couldn't be used on WSN[2]. WSNs also suffer with high failure rates. Further more, sensors don't have a high processing power. They hold only a small amount of memory with less processing capability.

Under some circumstances such as noise, environmental interferences and obstacles sensors could produce biased reporting as well. So the field of WSN need more research to address different aspects of the design. As a vital aspect the protocols and design should provide means to save the precious energy of sensors. Thus, the routing algorithms should consume little amount of energy for its processing and provide alternate solutions to expand the life time. For instance some algorithms provide the opportunity to switch to sleep mode when the power reach critical stage[2]. The protocols should be simple but also fault tolerant to cope with failure situations, consider the computation capability for processing, manage the data to securely flow and ensure a better network performance.

The transport protocol plays a crucial role in the network performance of WSN. Transport protocol runs on top of network layer. Typically, transport protocol is responsible for segmentation, flow control, error control, error recovery and QoS. But, widely used TCP and UDP doesn't address energy saving and doesn't suit for WSN multi-hop networks[4]. WSNs require more capabilities. WSN has a heavy upstream traffic towards the sink in a narrow bandwidth. That could possibly produce congestion and packet loss. Further more, the wireless channel could pose packet loss due to path loss and fading[4]. So the transport protocol should cope with packet loss and congestion. In short, WSN should address end-to-end reliability and QoS(Quality of Service) in an energy saving way[4].

Primarily, again the transport protocol must give precedence to energy efficiency. As the sensors hold less energy transport protocol should ensure energy efficiency. In WSN packet loss is common due to congestion. Such packet loss will lead for retransmission of packets and this will dissipate the power. Subsequently retransmission distance becomes the next factor for energy loss. If a longer distance retransmission is used in transport protocol that would consume more power. Another factor that would affect energy handling is control messages[4]. Transport protocol exploits control messages for retransmission and recovery. That means more processing and more energy loss. So a small packet with shorter retransmission distance and few control messages will be less power hungry.

Simulation study

In a multi hop wireless sensor network nodes may experience higher rate of packet loss, routing instability and unfair prioritization. In a sensor network traffic flow the packets have to travel via several nodes to reach the source. The nodes in the middle have content with each other to get the priority. If a node in the middle is with a low contention packet it will inject more traffic to get highest priority. But that might result an unaffordable traffic to the following nodes which have to forward. So the traffic has a high probability of packet loss and produce routing instability with retransmission. Further, the traffic with less contention gets high priority by injecting and cause inequality with traffic flow[5].

The entire simulation is carried on NS2 environment. The initial node or the source node will be node_0 and M stands for number of number of hops between source and destination. AODV protocol is specified and the distance between nodes is 200m. After the design generated a UDP agent is added to node_0, that is the source and CBR traffic (Constant Bit Rate) is generated to flow. The destination node will also have a UDP agent to sink the traffic. The 802.11 MAC type with half duplex link exists between nodes. Transmission range was 250m and Carrier-sense range was 550m.

The CBR packet defined as follows, CBR packet Size 2048, UDP packet Size 2048, CBR interval 0.0008, CBR random 0.96749, CBR maxpkts 1000000.

This simulation executed with different number of nodes. The following command used to run the simulations,

Ns multi-hop.tcl - 2{number of nodes}

In the next step the TR file was analysed to obtain the number of sent packets, received packets and dropped packets.

Sent packets - grap "^s" | WC -l

Received Packets - grap "^r" | WC -l

Drop Packets - grap "^s" | WC -l


The multi-hop sensor network is analysed for the performance. To derive conclusion on performance the deciding factors throughput, packet loss ratio and delay have been considered. The following equations have been applied to measure those factors.

Packet Delivery Ratio (PDR) = n of Received Packets / n of Sent Packets * 100

Packet Loss Ratio = Packets Dropped /Packets Sent *100

Throughput in bits/Second (bps) = Received Packet* Packet size * 8 / Time

The Fig. 1 depicts how the throughput varies with the increasing number of intermediate nodes. As expected, the throughput of the network drops down when the number of intermediate nodes increase.


Fig. 2 illustrates the Delivery ratio against the number of nodes. So when number of intermediate node increase delivery ratio also goes down.

Fig. 2

Fig.3 depicts the number of packets dropped when the in-between node numbers increased. It explains that packet loss increase when the packets travel via several nodes as explained in theory.

Following table provides the values derived via the simulation.


Sent pkts.

Received pkts.

Dropped pkts.





































Throughput will contribute to network performance evaluation. Throughput portrays the average rate of successful delivery. From the study we could verify that, when packets travel via less intermediate nodes performance will be retained. In other words for a better WSN design the performance factor should be addressed by keeping less number of intermediate nodes .