Optimization of the storage database for the Monitoring system of the CTA

TL;DR

This paper presents preliminary results on sizing and optimizing the hardware for the CTA monitoring system database, focusing on maximizing data write performance for large-scale time series data.

Contribution

It introduces an optimized configuration for the database hardware to handle big data time series in the CTA monitoring system, validated through testing.

Findings

Optimized database configuration increases data writing speed.

Feasibility of hardware sizing for CTA monitoring system demonstrated.

Test results support deployment of the monitoring database.

Abstract

We present preliminary test results for the correct sizing of the bare metal hardware that will host the database of the Monitoring system (MON) for the Cherenkov Telescope Array (CTA). The MON is the subsystem of the Array Control and Data Acquisition System (ACADA) that is responsible for monitoring and logging the overall CTA array. It acquires and stores monitoring points and logging information from the array elements, at each of the CTA sites. MON is designed and built in order to deal with big data time series, and exploits some of the currently most advanced technologies in the fields of databases and Internet of Things (IoT). To dimension the bare metal hardware required by the monitoring system (excluding the logging), we performed the test campaign that is discussed in this paper. We discuss here the best set of parameters and the optimized configuration to maximize the database data writing in terms of the number of updated rows per second. We also demonstrate the feasibility of our approach in the frame of the CTA requirements.