Network Management Deployment Plan - Research Proposal Example

Network Management Deployment Plan Optimal Resources Management in Wireless Sensor Networks [Pick the Contents 3 Introduction 3Network Management 5 Network Deployment Architecture 6 Proposed Methodology 7 Advantages of the Proposed Method 10 Limitations 11 Conclusions 12 References 13 Abstract This research paper discusses the intelligent deployment of the network resources to maximize their performance with given capabilities. This study considers a special type of network called as Wireless Sensor Network, comprising sensor nodes for deployment in the field of interest. These sensors are limited in their capabilities and resources. It is required to optimize the usage of their resources for maximizing the network performance. We will not go deep into networking protocols, as our main concern will be at application level of networking only. This study will optimize the network performance with the objective being the minimization of the usage of resources. We will develop an objective function that need to be minimized to achieve our research goal. Introduction Network is the combination of various elements that are used to create a channel for communication between two or more points. Plenty of networks exist all around us such as mobile networks, wired PTCL networks, cable networks etc. Network management is another specialized art that requires the manipulation of resources in intelligent manner to enhance the network performance and increase the customer satisfaction. This research paper is dedicated to a special type of network called as Wireless Sensor Network (WSN) that is built from a combination of sensor nodes (wireless sensors), deployed in the field for measuring and controlling some physical phenomenon. These sensor nodes contain some specific types of sensors embedded on them such as temperature, pressure, humidity etc. These sensor nodes measure the desired physical phenomenon and relay the information back to the base station. The information is collected through routing in the form of a network. Sensor nodes keep on relaying information from one hop to another until the information reaches base station. Although, extensive research has already been carried out in the field of network resource management in Wireless Sensor Networks (WSNs) in the past by researchers, yet the minimization of resources from the application point of view is still an unaddressed issue. Sensor nodes are always sensing the physical phenomenon, and their battery keeps on depleting. Network management requires making intelligent decisions regarding monitoring schedules of sensor nodes so that they don’t get out of fuel so early. These sensor nodes have to stay in the field for a long time since they are mostly used to relay information from inaccessible places. The applications of WSNs are plenty in industry today such as surveillance in battle fields, ecosystems etc. It is quite important to manage the network for utilizing their potential to maximum. The major issue in practicability of the WSNs is their short life span due to small battery contained in the sensor nodes. Owing to this fact, sensor nodes are mostly programmed to have small awake periods only when required. They go to sleep state when there is no application is to be executed. The base station sends the signal for making particular sensor node awake. Sensor node senses the phenomenon in a given time span, and relays the information back to the base station, and goes to sleep again. Network Management The basic purpose of the network management is to monitor and control the elements of the network to improve and sustain their performance. You will find many traditional network management systems for WSNs such as Manna (Ruiz et al, 2003) and Boss (Song et al, 2005). The former provides policy based management of sensor nodes while the latter is concerned with the UpnP networks and sensor nodes. Some novel routing protocols have also been proposed in the past regarding network management such as RRP (Liu et al, 2004) uses the zone flooding protocol, WinMS (Lee et al, 2006) works on the FlexiMAC protocol, and SNMS (Tolle & Culler, 2005) utilizes the Drip protocol. None of the aforementioned protocols address the optimal utilization of the network resources to enhance the lifetime of the WSN. The motivation of the topic comes from the fact that WSNs comprise highly expensive and sensitive sensor nodes that need large cost for their deployment. Hence their usage should be made quite intelligently. Their battery life is quite crucial. Network management makes sure that these sensor nodes never run out of resources. Network management is monitoring and control of all the process going on in WSNs for sensing, gathering and relaying information. In our proposed method, we are concerned with the application part of the sensor nodes while ignoring the relaying part that would make things more complex. We will assume that minimization is achieved to the desired extent such that even if communication takes more battery power, the overall life is longer than before with considerable difference. Our analysis will be composed on theoretical framework without employing any practical examples. This research will be a great step towards establishing the theoretical background of resource optimization of sensor nodes in WSNs. Network Deployment Architecture Wireless Sensor Network is usually deployed with the combination of various sensor nodes having the ability to communicate with the central sever called as base station. This central hub receives the information from all the sensor nodes, and passes it to the controller that takes action according to the requirements, and sends command to the base station that in turn issues these commands to the sensor nodes. This process continues in an infinite loop. The communication between sensor nodes and their talk with the base station is made through wireless mode of communication using radio frequencies. Sensor nodes may seem to be tiny but they have complete built in module of transmitter and receiver that can talk with other sensor nodes and base station. The radio frequency range of a sensor node will comprise its region of influence. Considering the range of sensor nodes to be circular, we will have many circles surrounding sensor nodes indicating their region of communication. Sensor nodes may be grouped according to their maximum common region of influence. This is one of the most popular sensor groupings also called as clustering. Each cluster of sensor nodes comprises one hop that may be called as a distinct step in communication. Multiple hops keep on transmitting the information until it reaches the base station. It is not necessary that each node in the cluster takes part in communication with the next immediate hop. Only one powerful node may be specified for this purpose that will transmit the information to the next cluster and so on. The communicating node may be callas Cluster Head. The main focus of this research is to utilize the resources of sensor nodes in intelligent way so that they can stay in the field for considerably long periods of time. Their function will be analyzed carefully that will tell about their requirements of processing, sensing, and communicating. Choosing cluster head for communication is one of the steps in making optimal usage of resources. Rest of the management will require adjusting the power for processing, storage, conversion into various forms, communicating when required. The tasks performed by the sensor nodes require consumption of battery a great deal. This battery consumption can be minimized by invoking the use of sensor nodes only when required. They should be allowed to sleep at all other times. Proposed Methodology Our research effort will comprise building an objective function that will require minimization for achieving proper results. Consider the following objective function needed to be minimized as follows. Where x represents the resources of the sensor nodes such as battery power, processing capabilities, neighborhood etc. “A’s” represent the weight of the resources depending upon their importance in the given scenario. In a given situation, the values of the “A’s” remain fixed while we have to play with the variable “x” that will minimize our objective function. The value of the constant parameters is determined from importance of the given parameter. For example, some tasks require only sensing without any processing capabilities or transmission to other sensor nodes. In such cases, the constant parameter related to the battery power will be high while for other resources, it will be low. It is obvious that the effect of that variable will be pronounced in the objective function that has greatest weight, and the minimization can be achieved based on the values of that resource. We may call the given resource that is critical in a given situation as critical resource. It is not necessary that we only have one critical resource; we may have a combination of critical resources (more than one ‘x’). In such cases, the objective will be to minimize the consumption of these resources as much as possible. To achieve minimization, we use numerical algorithms in PC since manual calculations will not yield suitable results. Numerical calculations will be performed in the master computer that is present at the base station while the results will be applied to the sensor nodes deployed in the field. In order to achieve suitable results practically, we should consider the importance of each resource individually and as a combined objective function. As an individual resource, their values represented by “x” should stay nearly the same with the course of the time. It means that the gradient of the objective function with respect to the critical parameters should be equal to zero. As a complete objective function with several critical parameters, we can take partial derivative of the objective function with respect to the available critical resources, and then find the minimization value that serves our purpose. The biggest question arises how do we control the network life from the value of the objective function? Well, it is not that simple to practically implement this algorithm since it may involve some complications. First of all, we need to find the values of the critical parameters in the given application, and then these values are utilized in the programming of the sensor nodes for the same application. Once the application is over, the objective function needs to be re-evaluated. The problem will be more complex once sensor nodes are running multiple applications with different requirement of resources. In such cases, the objective function could take the following form. Where “f’s” show the individual objective function for each application in a given scenario while total objective function is the sum of the individual objective functions. In order to implement the said strategy we may propose the following pseudo-code as follows. “Start, Receive command from user through base station, Find appropriate values of A’s in objective function, Take derivative of objective function equal to zero, Find minimization values of the resources, Program sensor nodes to achieve the minimization value, End” Advantages of the Proposed Method This method is clearly aimed at minimizing the consumption of resources for sensor nodes that will increase their lifetime, and hence saving the cost of deploying the network infrastructure. This is perhaps the best possible approach of network management in critical networks as WSNs. This method works well when the objective function is small with small number of critical resources to be minimized. The applications of the proposed methods are great in the industries. The sensors are usually used for sensing and relaying purposes in the field that barely requires some processing, and sensing powers. So their battery consumption should be kept to the minimum. With the aid of the said strategy, we can easily control the network traffic. Even with multiple applications, the sensor nodes will be programmed accordingly to adjust their critical resources according to the requirements. The proposed method is unique and innovative in the sense that special consideration is given to minimize the consumption of network resources that would be really helpful for the network in the long run. That is why we may assign some suitable name such as “Resource-based Network Management Protocol (RNMP)” to it. This protocol lies in strong competition with the proposed network management techniques and protocols. It will certainly enhance the network life from the applications point of view. The said strategy is very useful from the implementation point of view. There is not a great deal of computational burden on sensor nodes. Only the base station has to perform objective function minimization that itself is a master PC with high processing capabilities. Limitations While each method has advantages, it also has some associated limitations imposed by the nature, and scenario in which it is applied. When you consider the objective function only, some limitations are obvious. The objective function seems to have simplified the things related to communication between sensor nodes. In reality, the resources will be utilized in communication in addition to the application execution from the user point of view. Hence, the resources may be wasted in improper routing or communication method that would waste our efforts to optimize the resource utilization. Another limitation is imposed by the practicability of the method. It is difficult to update the programming codes of the sensor nodes at the run time that requires some attention at the base station. In order to make this possible, you need to input the programming codes of all possible applications in different scenarios, and adjust them so that the required code is programmed as desired in the given scenario. This task becomes increasingly difficult once more than one application are to be executed on the network, and you need to apply the combination of codes for sensor nodes. The proposed method doesn’t include the possibilities of the sensor node failure that is quite probable if it gets depleted of its battery power. In such cases, objective function in our case will stay the same that is not a good solution to the problem. There should be adaptable objective function with variable pack of resources instead of fixed whose values keep on decreasing with time. The proposed method doesn’t utilize the capabilities of the Cluster Head in routing information intelligently that would also increase the lifetime of the network. The purpose of Cluster Head is completely ignored. It may have been incorporated in the objective function to minimize it even further so that intelligent routing could be performed in addition to the utilization of other resources intelligently. Conclusions The proposed method considers the intelligent usage of the Wireless Sensor Network resources so as to enhance the lifetime in purely theoretical manner. For this purpose, we have proposed an objective function from the point of view of application that is to be minimized. Once the minimum value is attained, the minimization variables are programmed into the sensor nodes through base station that would be sufficient for ensuring the desired increase in network life. Our proposed algorithm also works for more than one application to be executed on WSN at a time. For this purpose, we have modified the objective function to include more than one objective functions dedicated to each application. Now minimization of the overall objective function is desired that will achieve our resource management aim. The minimization variable are again coded as desired by the application into sensor nodes through base station and this process continues forever. References Ruiz, L. B., Nogueira, J. M., & Loureiro, A. A. (2003). Manna: A management architecture for wireless sensor networks. Communications Magazine, IEEE, 41(2), 116-125. Lee, W. L., Datta, A., & Cardell-Oliver, R. (2006). WinMS: wireless sensor network-management system, an adaptive policy-based management for wireless sensor networks. Liu, W., Zhang, Y., Lou, W., & Fang, Y. (2004, October). Managing wireless sensor networks with supply chain strategy. In Quality of Service in Heterogeneous Wired/Wireless Networks, 2004. QSHINE 2004. First International Conference on (pp. 59-66). IEEE. Tolle, G., & Culler, D. E. (2005, January). Design of an application-cooperative management system for wireless sensor networks. In EWSN(Vol. 5, pp. 121-132). Song, H., Kim, D., Lee, K., & Sung, J. (2005, April). UPnP-based sensor network management architecture. In Proc. ICMU Conf. Read More