A novel data reduction technique with fault-tolerance for internet-of-things
Wong Siaw Ling, Boon-Yaik Ooi, Liew Soung Yue · 2017
Data reduction or data filtering using IoT-gateway is now on the rise in IoT research to facilitate the huge increase of data generated from IoT deployment, and solving problems such as network congestion, limitation of storage and bandwidth. However, existing solutions fall short in two problems, firstly lack of high availability concept to solve single point of failure (SPOF). Secondly, there is not state-of-the-art high availability solution that can be deployed without significant increase of cost and duplication of data. Therefore, this paper will present a concept that utilizing redundant IoT-gateway to improve the availability of data reduction by dividing the task among the IoT-gateways, without generating or sending any duplicated data to the defined end-point (e.g. database in the cloud). The concept is validated with our results that showing in the worst case, we can still achieve 76% accuracy by forwarding only 10% of data.