Performance evaluation and energy consumption of a video surveillance system with distributed storage

Abstract

Video monitoring systems are widely used as a tool for access control, crime and accident prevention, and internal security in various establishments. The storage of the data and images generated by these systems is a major challenge, as the number of devices, the quality of the images, and the required storage time have impacted the use of hardware resources and, consequently, electricity consumption. Evaluating the relationship between the number of cameras and the storage methods is crucial in proposing strategies that ensure reliability and availability at the lowest possible cost. It is important to verify the criticality levels of the images to be stored to determine which storage configuration will be used. The results indicate that the storage method (distributed or replicated volumes) has distinct impacts on hardware usage and electricity consumption. For example, it was possible to observe that the use of replicated volumes increases electrical consumption by about 6.17% with 100 cameras, compared to the distributed mode, and that the greatest impact on performance is in the utilization of the video management system CPU, independent of the storage mode.