The evolutionary advancements in the field of technology have led to the instigation of cloud computing. The Internet of Things paradigm stimulated the extensive use of sensors distributed across the network edges. The cloud datacenters are assigned the responsibility for processing the collected sensor data. Recently, fog computing was conceptuated as a solution for the overwhelmed narrow bandwidth. The fog acts as a complementary layer that interplays with the cloud and edge computing layers, for processing the data streams. The fog paradigm, as any distributed paradigm, has its set of inherent challenges. The fog environment necessitates the development of management platforms that effectuates the orchestration of fog entities. Owing to the plenitude of research efforts directed toward these issues in a relatively young field, there is a need to organize the different research works. In this study, we provide a compendious review of the research approaches in the domain, with special emphasis on the approaches for orchestration and propose a multilevel taxonomy to classify the existing research. The study also highlights the application realms of fog computing and delineates the open research challenges in the domain.