Abstract
The Internet of Spatial Things (IoST), continuously generates a large volume of geospatial data from a large number of connected smart devices. Cloud computing is inefficient to respond to IoST because of latency concern. Fog computing allows geospatial data processing as well as geospatial storage close to IoST devices and addresses this concern. This fog computing-based GIS model i.e. FogGIS is assisting to the cloud for invoking various geospatial web services. In many scenarios the service requests arrive to the fog layer and waits in a queue due to the current workload of fog servers. When all the Virtual Machines (VMs) are congested (busy), the arrived tasks are queued in the fog buffer until the fog servers becomes available to process the service. This creates the possibilities of task reneging or dropping. Reneged or dropped tasks could be resubmitted if the tasks were not dropped due to security reasons. In this paper, it experiments the performance of the fog layer based on the task arrival and the buffer size of the system using M/M/c/N Queuing theory. It also offers numerical illustrations to demonstrate the application of task reneging and feedback on the parameters such as queuing delay, probability of immediate service as well as probability of task rejection.