The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLII-4/W2
https://doi.org/10.5194/isprs-archives-XLII-4-W2-155-2017
https://doi.org/10.5194/isprs-archives-XLII-4-W2-155-2017
05 Jul 2017
 | 05 Jul 2017

EXTENDING THE SCALABILITY OF ISTSOS WITHIN THE 4ONSE PROJECT

D. Strigaro, M. Cannata, M. Cardoso, M. Antonovic, and M. Hoffmann

Keywords: istSOS, istSOS-proxy, EMS, WNS, Big data, SOS, scalability, 4onse

Abstract. An Environmental Monitoring System (EMS) is needed not only to prevent many natural risks such as droughts, flooding and landslides but also to provide information for a better management of water resources and crops irrigation and finally it helps to increase the reliability of weather and climatological models. In addition, a monitoring system can directly impact the economic, social and political spheres. Unfortunately, in most developing and low income countries, due to the high costs of hardware and software there is a lack of efficient monitoring systems. The aim of the 4onse project (analysis of Four times Open Non-conventional system for Sensing the Environment), funded by the Swiss National Science Foundation, is the development of a totally open solution to monitor the environment. As well as the hardware layer, a monitoring system needs a data management software usually hosted on a server structure. As a software platform which is SOS OGC compliant, istSOS is chosen to receive, manage, validate and distribute environmental data. In the following article a solution to support big data is presented to extend the istSOS capabilities. In fact, a sensors network can hardly stress a data management system because of the several concurrent users and sensors and the long time series which every weather station can easily produce. Thus, a software called istSOS-proxy is developed as a single access point over multiple instances of istSOS whose procedures are distributed to balance the total load. First results on the effectiveness of the solution are proved thanks to load testing simulations of different levels of concurrent users.