Experimental Evaluation of Publish/Subscribe-Based Spatio-Temporal Contents Management on Geo-Centric Information Platform

Abstract

Cross-domain data fusion is becoming a key driver to growth of the numerous and diverse applications in IoT era. Nevertheless, IoT data obtained by individual devices are blindly transmitted to cloud servers. We here focus on that the IoT data which are suitable for cross-domain data fusion, tend to be generated in the proximity, and thus propose a Geo-Centric Information Platform (GCIP) for the management of Spatio-Temporal Contents (STCs) generated through the cross-domain data fusion. GCIP enables to keep STCs near the users (at an edge server). In this paper, we practically examine the fundamental functions of the GCIP from two aspects: (1) Geo-location aware data collection and (2) Publish/Subscribe-based STC production. Furthermore, we implement a proof-of-concepts (PoC) of GCIP and conduct experiments on a real IPv6 network built on our campus network. In this experiment, we showed that multiple types of IoT data generated in the proximity can be collected on the edge server and then a STC can be produced by exploiting the collected IoT data. Moreover, we demonstrated that the Publish/Subscribe model has a potential to be effective for STC management.