Literature Review on Wireless Sensor Networks
The Micro-electromechanical system, processor, radio, and memory technologies development have made it a possibility to develop micro sensor nodes. These nodes are advantageous in the sense that they are relatively cheap, small in size and use small amounts of power in sensing, wireless communication, and computation. As such, it can be concluded that the sensor network is as a result of a combination of distributed information techniques, sensor techniques and communication and processing techniques. When thousands or hundreds of these nodes are spread in a large geographical area, they form a Wireless sensor network (WSN). The ambient environmental conditions surrounding these sensors are measured and the data is transformed in the form of electrical signals that undergo processing, revealing various phenomenon characteristics in the reference area. Every sensor in the network can collect and communicate data to other sensors or to the external base station (Shen & Zhao, 2012). In this case, the base station may be a mobile node or a fixed node that allows for the connection of the sensor network to the internet or an infrastructure of communication, from which a user can access the recorded data. As such, a user is able to access information concerning an area that is far from their physical reach. Some of the common applications of this technology include robot control, office buildings environment control, guidance in environments of automatic control, and smart homes using high security, among other numerous uses.
The routing process in sensor networks poses a great challenge as these sensors present various characteristics that are distinct from the modern wireless Mobile Ad-hoc Networks (MANET). One of the most important differences between these networks is that the WSN sensor nodes are higher in magnitude as compared to those in MANET (Shen & Zhao, 2012). On the other hand, the sensor nodes in WSN are less likely to have an ID like those in MANET. It is also evident that the sensor nodes are low in price as compared to the nodes in a MANET, and the former are normally deployed in high numbers. Nevertheless, the nodes in a MANET can be recharged whereas the sensor nodes have very limited resources of power. The low cost of the sensor nodes comes at a price, as they tend to be limited in terms of their communication and computation capabilities in comparison to MANET. In addition, the sensor networks are more likely to experience failure due to their frequent changes in topology. Broadcasting is the major communication paradigm used by sensor nodes, while the Ad-hoc networks utilize point-to-point communication (Shen & Zhao, 2012). Such differences have led to the development of various new algorithms in the sensor networks for routing data. As such, there is need to study the WSN routing protocols. Both the industrial and academic world have recently shown increased interest in the WSNs, with their efforts being directed towards issues related to the development of low cost, energy efficient, fault-tolerant ad secure sensor networks.
Classification of Routing Protocols in WSN
Various ways of classifying routing protocols in the WSNs have been suggested. Nevertheless, hierarchical, data-centric, location based platforms form the base for classification of such protocols in accordance with the structure of the network. All nodes in the data-centric routing are assigned typically equal functionality or roles. However, the nodes in the hierarchical-based routing have different roles to play in the network. On the other hand, the positions of the sensor nodes in location-based routing are exploited in the routing of data in the network.
Data-centric Routing Protocols
Instead of individual nodes, this paradigm allows for the combination of various applications that are required to access data, with a natural network-processing framework. The random deployment of sensor nodes coupled with unavailability of a global recognition, makes it difficult to select the appropriate sensor node sets that are to be queried. As a result of such a consideration, the data-centric routing has been developed. This type of routing is different as compared to the tradition address-based routing, which involves the creation of routes between nodes that are addressable. The Sensor Protocols for Information via Negotiation (SPIN) data-centric protocol is the first protocol and promotes the negotiation of data between nodes with the aim of eliminating redundant data and saving energy (Singh, Singh, & Singh, 2010). On the other hand, the development of Directed Diffusion has defined a new approach in data-centric routing.
SPIN is one of the early approaches of data-centric routing. The major feature of this routing protocol, which makes it ideal in large-scale use, is that it utilizes meta-data in the naming of data through a description of the data’s characteristics. As such, the protocol involves sending of three message types, including ADV, REQ, and DATA (Jiang & Manivannan, 2004). In this case, the sensor node advertises any DATA message before sending it through a broadcast of the ADV message, which comprises of a description of the data or meta-data. In the case that a neighboring sensor node is interested in the message, it responds by sending a REQ message, which includes a request for the DATA. It is at this point that the DATA message comprising of the actual data is send to the neighboring node. This process is repeated by the neighbor node until the data reaches the base station. One of the major advantages of this system that makes it a consideration in the monitoring of boarders is the fact that there is localization of topological changes, as each node is only required to recognize its single-hop neighbors. As such, in a global environment, where it is difficult to identify a certain set of nodes from which data is to be retrieved, this mechanism allows for ease in selecting the specific nodes from which data can be accepted. This can be done by identifying the title and description of the data in the meta-data sent through the ADV message before accepting such data (Jiang & Manivannan, 2004). This allows for the transferability of data that is only relevant to a network and the ignoring or data that is not helpful to the user.
Nevertheless, the SPIN data-centric routing protocol presents with various disadvantages, which deem it less applicable in large-scale environments such as border monitoring. One of the disadvantages is that the protocol is not scalable (Jiang & Manivannan, 2004). Thus, there is no limit to the shared data and as much as redundancy is eliminated, classified messages can be communicated to a different node and give away access to unintended users. Secondly, the nodes have low energy efficiency. Thus, in a case where the sink has interest I many events, the nodes around it are likely to deplete their energy. This compromises the sustainability of the system. Given the sensitive nature of border monitoring, such system downtime can lead to unnoticed occurrence of various events that the same system has been developed to monitor and communicate. Lastly, the mechanism involving the advertisement of data prior to sending does not guarantee if the data has been delivered (Jiang & Manivannan, 2004). As such, in the cases where there is communication breakdown, it may be difficult to realize, as it is expected for nodes to accept data that they are only interested in, and in cases where they do not accept, then an assumption is made that they are not interested in the involved data. In some cases, the data may be important yet it fails to be delivered. For instance, in some cases, the node that is interested in given data may be far from the node that has the data of interest. However, in the case where the nodes between the two communicating nodes are not interested in the data, then the interested node will not receive the data (Felemban, 2013).
Hierarchical Routing Protocols
Hierarchical routing sensor nodes are similar to cellular telephone network in a way that they transmit data to a centralized cluster head, from where the data is further distributed to the appropriate recipients. As such, this routing’s major aim is to reduce the consumption of energy by sensor nodes through involving them in a multi-hop form of communication inside a cluster and through promoting data fusion and aggregation with the goal of reducing the number of messages send from the nodes to the base station. This feature of energy efficiency is one that makes it an appropriate choice for large-scale application. Among others, PEGASIS and LEACH are examples of hierarchical routing protocols (Kour, 2012).
Low Energy Adaptive Clustering Hierarchy (LEACH)
In the LEACH routing protocol, a cluster-based approach is used, where the head of a cluster receives data from different sensor nodes within the cluster, and transfers the data to the base station upon aggregation. The algorithm randomly conducts changes in terms of the head of the cluster, bringing in a new head, which consumes more energy as compared to the other nodes in the cluster each time. This is meant to draw equity in terms of energy use among the nodes of the cluster, thus extending the networks lifetime. The head of the cluster conducts data aggregation and sends it to the base station, an aspect that reduces the cost of communication. Each node is required to select a number between zero and one. A node that gets a number below the following threshold assumes the role of the head for the given round (Kour, 2012).
The rounds in the LEACH protocol start with the set-up phase and then they enter into the steady-state phase. It is at the set-up phase that each node that is not the head of the cluster communicates its decision to the head of the cluster the CSMA MAC protocol (Kour, 2012). The heads of the cluster then develop and broadcast TDMA schedules to the cluster members during the steady state phase. Each node waits for an opportunity to transmit required data during the phase of data transmission. It could be argued that due to the energy efficiency of LEACH, where it reduces energy use by seven factors as compared to the other means of direct communication, it forms a better protocol in large-scale environments such as border monitoring, which require fewer disturbances in terms of energy rekindling. In addition, long life of the system resulting from dynamic clustering is an important feature of this protocol, as it allows for system sustainability and allows for maximum surveillance without breaking. More bovver, this routing protocol does not require global network knowledge as it is completely distributed. However, the major feature that distinguishes LEACH as a routing protocol, which is the use of a single-hop routing that allows for the transmission of data directly from every node to the head of the cluster and the base station, makes it inapplicable in large-scale networks (Kour, 2012). In addition, an extra overhead is brought about by dynamic clustering. For instance, changes in the heads and advertisements among other factors are likely to decrease the gain in consumption of energy.