INTEGRATING EDGE COMPUTING FOR ENHANCED RELIABILITY IN DISTRIBUTED VIDEO SURVEILLANCE SYSTEMS

Distributed video surveillance systems have significant communication overhead, are dependent on cloud services, have latency in analytics and are prone to service outages due to network failures. In order to overcome the aforementioned challenges, this paper proposes an edge-computing based framework to improve the reliability, scalability and responsiveness of distributed surveillance environments. The proposed framework combines local processing of the video stream, real-time edge analytics, fault-tolerant operation, and adaptive load balancing between the edge nodes and the cloud infrastructure. The framework processes video streams locally at the data source, decreasing the communication overhead, the dependence on the cloud, and making it possible to monitor even in the event of a network outage. Moreover, the architecture enables efficient use of resources by dynamically distributing workloads and providing robust edge intelligence. Experimental assessments showed significant performance gains compared to traditional cloud-based surveillance solutions, such as a 42.8% decrease in analytics latency, a 48.6% decrease in network bandwidth usage and a 97.1% service availability. The framework persisted to have solid monitoring and effective resources utilization even in the face of bad network conditions. The proposed framework brings together the capability of fault-tolerant edge intelligence and distributed video analytics, ensuring a reliable and scalable framework for large-scale IoT-based surveillance systems with the need for continuous monitoring, efficient resource management, and resilient operation.