R&D: Derived Multi-Objective Function for Latency Sensitive-based Cloud Object Storage System using Hybrid Heuristic Algorithm

Data & Knowledge Engineering has published an article written by N Nataraj, assistant professor, Bannari Amman Institute of Technology Sathyamangalam, Tamil Nadu 638401, India, and RV Nataraj, professor, Erode Sengunthar Engineering College Thuduppathi, Erode, Tamil Nadu 638057, India.

Abstract: “Cloud Object Storage System (COSS) is capable of storing and retrieving a ton of unstructured data items called objects which act as a core cloud service for contemporary web-based applications. While sharing the data among different parties, privacy preservation becomes challenging. Research Problem: From day-to-day activities, a high volume of requests are served daily thus, it leads to cause the latency issues. In a cloud storage system, the adaption of a holistic approach helps the user to identify sensitive information and analyze the unwanted files/data. With evolving of Internet of Things (IoT) applications are latency-sensitive, which does not function well with these new ideas and platforms that are available today. Overall Purpose of the Study: Therefore, a novel latency-aware COSS is implemented with the aid of multi-objective functionalities to allocate and reallocate data efficiently in order to sustain the storage process in the cloud environment. Design of the Study: This goal is accomplished by implementing a hybrid meta-heuristic approach with the integration of the Mother Optimization Algorithm (MOA) with Dolphin Swarm Optimization (DSO) algorithm. The implemented hybrid optimization algorithm is called the Hybrid Dolphin Swarm-based Mother Optimization Algorithm (HDS-MOA). The HDS-MOA considers the objective function by considering constraints like throughput, latency, resource usage, and active servers during the data allocation process. While considering data reallocation process, the developed HDS-MOA algorithm is also performed by considering the multi-objective constraints like cost, makespan, and energy. The diverse experimental test is conducted to prove its effectiveness by comparing it with other existing methods for storing data efficiently across cloud networks. Major findings of results: In the configuration 3, the proposed HDS-MOA attains 31.11%, 55.71%, 55.71%, and 68.21% enhanced than the OSSperf, queuing theory, scheduling technique, and Monte Carlo-PSO based on the latency analysis. Overview of Interpretations and Conclusions: The developed HDS-MOA assured the better performance on the data is preserved in the optimal locations having appropriate access time and less latency that is highly essential for the cloud object storage. This supports to enhance the overall user experience by boosting the data retrieval. Limitations of this Study with Solutions: The ability of the proposed algorithm needs to enhance on balancing the multiple objectives such as performance, cost, and fault tolerance for optimally performing the operations in real-time that makes the system to be more efficient as well as responsive in the dynamic variations in the demand.“