Cognitive IoT Gateways - overview
With the proliferation of IoT devices and sensors on the Internet into the billions, the assumption that all computation can simply be pushed into the Cloud is not practical. Edge/Fog computing is needed to help Cloud computing serve more IoT applications efficiently. Currently, some applications are assigned to be served by Edge/Fog, some by Cloud and the rest by hybrid Fog-Cloud. The platform for each application is predefined and unchangeable at run-time, regardless of the real-time conditions and the performance of the system part. The current IoT gateways are merely bridges between IoT devices and the Cloud and are not smart enough to decide when to run an application on the Fog/Edge platform and when to run it on the Cloud based on the system real-time conditions for achieving a better performance.
IBM scientists are working on a new research agenda to build dynamic switching between Fog/Edge and Cloud for IoT applications for improved performance and better utilization of Fog/Edge-Cloud capabilities. We are building cognitive IoT gateways that can learn and thus make dynamic real-time decisions to achieve optimal results. Such an intelligent gateway is an essential part of the exploding IoT ecosystem.