Abstract and subjects
The number of connected edge or internet-of-things (IoT) devices is expected to cross over 20 billion by the 2021. These devices include basic sensor nodes which generate data and upload them to the cloud via edge processing devices. While the speed of the data transfer has increased significantly, so has the amount of data that needs to be transferred. But most of the data collected at the edge is irrelevant for analysis and inferences, so transferring all the data is inefficient and expensive. The inferences made at the cloud need to be transferred back to the edge for execution, which can cause considerable lag in decision-making for real-time applications. However, the resource-limited environment of edge devices with limited computational capacity render them a challenging platform for the deployment of high computational data analytic and intelligence mechanisms, particularly for real-time applications. Data privacy is a crucial issue while the data from edge devices are being sent to the cloud for resource-intensive training of algorithms. Technological advancements have made it possible for embedded devices to communicate with IoT sensors in a simple and cost-effective approach using embedded software platforms, low-power and long range communication technologies. The embedded edge device acts as IoT gateways by collecting data at regular or predefined intervals from IoT sensors and pushes the data to the cloud. The cloud acts as a data center and performs most of the heavy lifting job in terms of building big data analytic to trigger important business decisions and transfer them to the edge for execution. This requires a distributed platform that integrates communication, computation, and storage resources for real-time applications. As the number of devices grows, achieving real-time and prioritized decision-making, control, and management in a centralized platform becomes challenging. Deploying intelligence-at-the-edge devices is also necessary along with maintaining a central cloud-based solution to enable agile connectivity, real-time control, and data optimization, ensuring data privacy and security. Hence, combining the computing power of cloud with cognizant and perceptive edge devices is essential to build an impactful business solution.