A Multi‐Layer Adaptive Resource Allocation Approach to enhance Resource Utilization for SDN‐Based Edge Computing

ABSTRACT

Today, with the rapid growth of the Internet of Things (IoT), the volume of data generated and processed has increased significantly, and there is an urgent need to handle tasks quickly and efficiently in real‐time applications. Existing cloud‐based models are not capable of meeting these demands due to their deployment model that adds latency. The separation of the control and data planes has the potential to provide a solution to Software‐Defined Networking (SDN) that offers centralised control, programmability, and dynamic network management. However, efficient scheduling of activities and workloads among the different edge servers is one of the imperative issues since an inefficient allocation of the resources may lead to server congestion, high latency rates, and resource wastages. To overcome this, the article presents the Average‐Based Load Balancing and Resource Allocation Mechanism (ALBRAM), which identifies suitable nodes and dynamically allocates resources to maintain the load balance among all the servers. It recommends a three‐layer SDN‐based edge computing architecture providing a bridge between the IoT devices, middle nodes, and the edge servers, considering both the communication and computation time. The strategy uses a least‐load server selection mechanism to provide fairness, optimal resource utilisation, and balanced workloads. The results of the evaluation indicate that ALBRAM decreases the Makespan, total completion time and latency and increases resource utilization, and the efficiency of load‐balancing in comparison to existing methods.