Abstract

There is a growing trend to move some of the analytics that are typically performed on cloud applications to the access side of the network of cyber-physical systems in order to reduce latency, improve response times and extend the infrastructure lifetime. However, this approach, known as mist computing, introduces problems associated with the broad sharing of sensor readouts. Specifically, smart devices are deployed in constrained networks and have limited power, making them constrained devices. Because these devices are unsupervised and rely on batteries for power, they are inherently power-limited. Moreover, to preserve battery life, they typically transmit data at very low rates. This situation leads to a number of issues in the context of smart device analytics. This paper introduces these problems and provides a solution that enables the efficient support of analytics in a network of constrained smart devices in the context of mist computing.

Statistics