Today's Internet of Things (IoT) is enabling innovations much faster to enhance the quality of life using various IoT applications such as Smart City, Smart Homes, Autonomous Driving Cars, Drone Monitoring Systems and many more applications. Cloud Computing in IoT has playing a significant role in providing its maximum services such as Data Storage, Computing or Processing, Analyzing and Securing Big Data etc for IoT applications. But, researchers estimated that the users are going to deploy 1 trillion IoT devices by the year 2025 which could have an economic impact of $11 trillion dollars per year. The unprecedented amount of data that is going to generate from those number of devices is pushing the cloud computing traditional systems to difficult situations to handle. Along with this situation there are certain applications (Self Driving, Health Monitoring, Gamming, Real-Time Big Data Analytics, Live Streaming and Controlling Applications etc.) in IoT which requires quick response to react and control the situations if necessary. The traditional Cloud Computing systems in this scenario cannot satisfy the service level agreement (SLA) made between the cloud service providers and cloud users. Balancing this scenario, cloud service providers must maintain a reasonable performance while reducing the energy consumption and cost of maintenance. In this thesis work a new computing paradigm called Fog Computing has been used for our newly proposed application models to overcome the problems or drawbacks of Cloud Computing and to enhance the performance for our application models by satisfying the quality of service level agreements (QOSLA). By using Fog Computing distributed strategy our models achieved better performance and worked more efficient than in Cloud. It also reduced the network usage, energy consumption in cloud and cost of execution. To simulate our work, we used iFogSim simulation toolkit which is currently growing significantly for the past couple of years.