Intelligent Resource Management Schemes in Vehicular cloud Networks

Abstract

newline Cloud computing integrated with Vehicular Ad Hoc Networks (VANET) is known as newlineVehicular Cloud Networks (VCN). The purpose of the VCN is to provide various ser- newlinevices to the user of the vehicle. Services may be related to the traffic information, or newlineservices may be related to infotainment, safety applications, nonsafety applications, loca- newlinetion awareness services, etc. Unpredictable mobility of vehicles is one of prime concerns newlinein VCN needs to be addressed as compared with the traditional network where the node newlineis usually. As VCN comprises of cloud and VANET technologies, various issues related newlineto the cloud and VANET environment need to be addressed to provide various services newlineto the vehicle at an appropriate time. newlinePresent vehicles consist of onboard equipment and sufficient resources and these re- newlinesources are underutilized. The Vehicular cloud consists of data centers and these data newlinecenters consist of huge amount of resources. These resources are scattered geographically newlineand need to be efficiently managed. Resource monitoring, discovery, allocation, sharing newlineetc, are some of the major issues in vehicular cloud computing. Privacy, security, virtual newlinemigration issues also need to be addressed for the vehicular cloud environment. newlineThe research work presented in the thesis resolve some of the above mentioned issues newlineto provide the services to the vehicle as and when required. The thesis attempts to newlineattain three research objectives to provide a better service (in terms of response time, newlinecost) to the user of the vehicle and also efficiently utilize the resources of a computing newlineunit (vehicle). newlineTo discover computing resources accessible for any application before they are allo- newlinecated to requests dynamically on-demand, design, and development of an effective mech- newlineanism for resource discovery in VCN is very important. Lack of intelligence methods newlinein resource discovery and less flexibility for simultaneous requests issues are the prime newlineconcerns for resource discovery in VCN. newlineIn the first objective, proposed dynamic resource discovery based on Honey Bee Op- newlinetimization (HBO). HBO technique is integrated with static and mobile agents. Mobile newlineagent collects the vehicular cloud information and static agent intelligently identifies the newlinerequired resources by the vehicle. Dynamic discovery model will take into account differ- newlineent parameters influencing the task execution time to optimize subsequent schedule. To newlinetest the performance effectiveness of the scheme, proposed dynamic resource discovery newlinescheme is compared with a fixed time scheduling algorithm. The objective of the pro- newlinei newlineposed scheme is to search the resources in VCN with a minimum delay. The simulation newlineresults prove that the proposed scheme is better than the existing scheme. newlineEfficient Resource management in Vehicular Cloud Networks (VCN) results in an newlineincrease resource utilization and reduction of the cost. An effective resource allocation newlinescheme in VCN plays a major role in the overall performance of the system. Members newlineof VCN change dynamically due to the mobility in their movement. Proper resource newlineallocation schemes in VCN provide the better performance in terms of the reduction of newlinecost, reduction in the waiting time of vehicle (client), and also the waiting queue length. newlineResources are required to provide more efficiently by the cloud providers for the requested newlineservices by the vehicle. For this reason, it is necessary to design proper resource allocation newlineschemes in VCN. newlineIn the second objective, the proposed resource allocation scheme in VCN allocates newlinethe appropriate computing resources for the client vehicle application. Vehicles may face newlinehigh costs or issues related to the performance parameter when proper resource allocation newlineschemes are not applied. The proposed cost model provides resources the to vehicle by newlineconsidering the lesser expensive approach by reducing the cost. We compare the results newlineof the cost optimization with the generic algorithm that uses a combination of best fit newlineand first fit techniques for resource allocation in VCN. newlineIn the third objective, we proposed a resource scheduling mechanism that schedules newlinethe tasks based on the priorities of the local tasks. Scheduler preempts the remote newlinetasks and executes the local tasks which are considered to be high priority. Scheduler newlineagent program running on computing engine (vehicle) handles task scheduling. Scheduler newlineplans execution along with local tasks, with the latter taking higher precedence. We newlinehave simulated the proposed resource scheduling scheme and compared the results with newlineexisting algorithm