Task scheduling and resource allocation of seasonal requests of users in cloud using NMKA and CM-GA techniques

Purpose The purpose of this paper is to provide energy-efficient task scheduling and resource allocation (RA) in cloud data centers (CDC). Design/methodology/approach Task scheduling and RA is proposed in this paper for cloud environment, which schedules the user’s seasonal requests and allocates resources in an optimized manner. The proposed study does the following operations: data collection, feature extraction, feature reduction and RA. Initially, the online streaming data of seasonal requests of multiple users were gathered. After that, the features are extracted based on user requests along with the cloud server, and the extracted features are lessened using modified principal component analysis. For RA, the split data of the user request is identified and that data is pre-processed by computing closed frequent itemset along with entropy values. After that, the user requests are scheduled using the normalized K-means algorithm (NKMA) centered on the entropy values. Finally, the apt resources are allotted to that scheduled task using the Cauchy mutation-genetic algorithm (CM-GA). The investigational outcomes exhibit that the proposed study outruns other existing algorithms in respect to response time, execution time, clustering accuracy, precision and recall. Findings The proposed NKMA and CM-GA technique’s performance is analyzed by comparing them with the existing techniques. The NKMA performance is analyzed with KMA and Fuzzy C-means regarding Prc (Precision), Rca (Recall), F ms (f measure), Acr (Accuracy)and Ct (Clustering Time). The performance is compared to about 500 numbers of tasks. For all tasks, the NKMA provides the highest values for Prc, Rca, Fms and Acr, takes the lowest time (Ct) for clustering the data. Then, the CM-GA optimization for optimally allocating the resource in the cloud is contrasted with the GA and particle swarm optimization with respect to Rt (Response Time), Pt (Process Time), Awt (Average Waiting Time), Atat (Average Turnaround Time), Lcy (Latency) and Tp (Throughput). For all number of tasks, the proposed CM-GA gives the lowest values for Rt, Pt, Awt, Atat and Lcy and also provides the highest values for Tp. So, from the results, it is known that the proposed technique for seasonal requests RA works well and the method optimally allocates the resources in the cloud. Originality/value The proposed approach provides energy-efficient task scheduling and RA and it paves the way for the development of effective CDC.