Cloud computing is an architecture which several Internet-of-Things (IoT) devices relay on to outsource their computation tasks. The numbers and needs of IoT devices have grown massively, to which cloud computing could not fulfill these needs. Cloud computing has its shortcomings in terms of bandwidth, latency, and real-time response. To overcome these shortcomings fog computing paradigm was introduced, which can fulfill requests from IoT devices and forward requests to the cloud if necessary. However, there are still requests that need to go to the cloud and might suffer from its shortcomings. Our proposed peer-to-peer (p2p) fog model enhances fog computing by adding p2p mechanism into the fog layer, which allow the fog nodes to collaborate in order to meet the clients needs. This way we minimize the requests that go to the cloud and fulfill most of the requests by the fog nodes at the proximity of the user. In this research, we evaluate the proposed p2p fog model, under the use case of file sharing application, and show that it has better outcomes in term of bandwidth throughput compared to cloud computing and fog computing models. Therefore we have simulated cloud computing, fog computing, and p2p fog computing scenarios, then compared and evaluated their outcome results. Further research into fog computing could build upon this proposed model and extend existing fog computing architectures.