Published Paper Details:

Cloud computing is a technology that utilizes a group of interconnected nodes to provide a range of services such as computation, data processing, and resource sharing. The nodes work together to deliver these services within a data center, with a master node serving as the coordinator for the nodes in a cluster. This master node is responsible for managing tasks such as handling deadlocks, reducing latency, ensuring reliability, improving performance, and providing users with access to efficient processing services. Selecting the optimal master node can be a complex task, as it involves considering factors such as memory, CPU-MIPS, throughput, and bandwidth. The current methods for choosing master nodes are generally inflexible, as they do not consider the resources available on individual nodes. To address this issue, a "watcher" program is employed to continuously monitor the master node's fitness for coordinating the network of nodes. If the master node is no longer suitable, the watcher program triggers a real-time reassignment process, during which the next most suitable node is appointed as the new master. This process occurs quickly and seamlessly, so that the user is not aware of any delay in response time. The Master Node Consensus Approach (MNCA) method proposed in this paper aims to enhance efficiency, minimize latency, and improve overall user experience. To achieve this, the Master Node Consensus Approach (MNCA) utilizes a genetic algorithm (GA) to select the optimal master node from the data center. The proposed approach consists of three modules: initial populace creation, master node appointment, and candidate node appointment. In the initial module, a scheduling algorithm is used to randomly assign tasks to different nodes/servers, generating the initial population. The overall performance and capabilities of each node are then evaluated based on available resources. In the second module, the best optimal nodes are identified as potential master nodes using genetic operations like crossover, mutation, and fitness functions, with available resources being taken into consideration. From the set of optimal nodes, one node is selected to be the master node, while the others are designated as candidate nodes. In the third module, a candidate node is assigned as the new master node in the event of the original master node's failure.