<p>The over-reliance of current methods for task allocation and resource management in distributed software-defined networks (DSDN) on constraints and predefined assumptions about network behavior leads to severe efficiency degradation in operational environments. This inefficiency is due to the dynamic and unpredictable nature of DSDN networks. Hampering assumptions, such as applying static task entry models, lead to suboptimal management and utilization in resource allocation and task distribution. To overcome these limitations, this study presents a novel approach called ELDM-EDSDN based on the effective improvement of dynamic mappings and manifold rules for resource management and allocation in Edge-DSDN networks for proper utilization of network resources and optimal task execution. The processing load status of controllers is initially monitored in real time. Once controllers with dense processing loads are identified, an adaptive threshold-based decision-making mechanism is activated. Then, additional tasks are transferred to less loaded controllers by considering criteria such as processing capacity, network latency, and topographic distance. Relying on minimal initial assumptions and the best possible strategy, this process achieves load balancing in the network and effectively manages resource allocation. Simulation of the proposed method on OMNET++ under various scenarios, including changing the number of controllers, task volume, and network load level, proves that ELDM-EDSDN improves the response time of controllers, network load balancing, and latency by 14.5, 21.17, and 14.64 percent, respectively, compared to peer methods. Also, the proposed method is less sensitive to changes in the number of controllers, servers, virtual machines, and input tasks, which is crucial in real dynamic networks.</p>