The traffic violations system is one of the systems those are based on the development of the internet and internet of things (IoT) as law enforcement systems. Such system is used to register the violations committed in the streets of the cities. According to the incremented numbers of vehicles per years, the traffic violations system requires an efficient network design to reduce latency and cost by distributing the edge server according to the spatial distribution of the loads in a network system. The service decentralization and the spatial clustering are used to distribute the service in the edge servers of the system by K-means as a machine learning algorithm and reduce the distances between nodes the number of the connected devices to each edge server. The results showed that the proposed system provides low latency for the suggested scenarios (50, 100, 250, 500 clients), compared to the standard central service, which were 400.05, 400.05, 400.05, and 800.1, respectively. It also proves the scalability of the system by adding more edge servers.