Towards a Framework for Data Stream Processing in the Fog

Abstract

In volatile data streams as encountered in the Internet of Things (IoT), the data volume to be processed changes permanently. Hence, to ensure timely data processing, there is a need to reconfigure the computational resources used for processing data streams. Up to now, mostly cloud-based computational resources have been utilized for this. However, cloud data centers are usually located far away from IoT data sources, which leads to an increase in latency since data needs to be sent from the data sources to the cloud and back. With the advent of fog computing, it is possible to perform data processing in the cloud as well as at the edge of the network, i. e., by exploiting the computational resources offered by networked devices. This leads to decreased latency and a lower communication overhead. Despite this, there is currently a lack of approaches to data stream processing which explicitly exploit the computational resources available in the fog.

Within this paper, we consider the usage of fog-based computational resources for the purposes of data stream processing in the IoT. For this, we introduce a representative application scenario in the field of Industry 4.0 and present a framework for stream processing in the fog.

Avatar
Stefan Schulte
Professor & MLE-Sprecher

Data Engineering, Internet of Things, Federated Learning, Blockchains