Current Edge Computing Solutions
In recent years, cameras are increasingly equipped with GPU or FPGA powered edge devices. This enables them to run deep learning algorithms for image processing while benefiting from numerous advantages.
Every type of event requires its own set of deep learning models. For example, traffic detection in an urban environment requires a different set of classifiers than crowd detention in public transportation. To circumvent problems such as computational limitation of edge devices, the few emerging solutions that exist, train a large number of models on the cloud and ship the trained models to the edge devices for execution.
erera is the first Opportunistic Software Defined Edge Computing solution for Smart Cities. By installing erera’s Edge OS on the city-owned edge devices, they can be pooled on-demand to create ad-hoc computing systems for executing and training the deep learning models independently of cloud infrastructure. This way, the city benefits from from computational capacity and SaaS solutions comparable to that of clouds while maintaining advantages of edge computing solutions.