A Cloud-Based Monitoring System Employing Frame Skipping and Path Switching

Abstract

A congestion-aware cloud-based monitoring system is presented. The proposed system is capable of continually accommodating its upload paths and adjusting its frame rate in response to network condition changes. Path switching scheme increases the throughput of cameras due to the avoidance of congested links. Frame skipping method eliminates the latency since cameras transmit only the frames which are likely to achieve their destinations in time. We further adopt an MPEG2-TS packer to encapsulate video and audio contents into MPEG2-TS files, which facilitate not only the distribution among several nearby data centers but also the instant or time-shifted playback of those distributed contents. The simulation results show that our solution yields a significant increase in the average throughput of all the cameras and achieves a lower latency whilst granting an acceptable responsiveness.