An Adaptive Resonance Theory-Driven AI Approach for Multi- Level Workload Classification in Autonomic Cloud Database and Data Warehouse Systems

Gopi Ramasamy; Bhagath Singh Jayaprakasam

Authors

Gopi Ramasamy Author
Bhagath Singh Jayaprakasam Author

Keywords:

Adaptive Resonance Theory (ART), multi-level workload classification, autonomous cloud systems, Artificial Intelligence-driven resource allocation 1 Introduction In a world of data-driven applications and rich interactive user interfaces that pull back-end, computations into the forefront via innovative technology stacks, long-running mixed transactional analytics processing workloads are characteristic for many web-scale environments like search engines or social networks

Abstract

Background Information: This work presents an AI-centric strategy for classifying multi-level workload in autonomic cloud database and data warehousing systems using Adaptive Resonance Theory (ART). The model is designed to better utilise resources, Improving system efficiency and allowing the handling of varied workloads on-the-fly. Objectives: This requires an ART-based AI model which classifies workloads correctly, optimises resource allocation in autonomic cloud systems, and guarantees adaptive learning of
dynamic workloads without abandoning previous knowledge. Methods: We categorized excellent solutions conducted here with ART-based classification and multi-level workload segmentation approaches, employing dynamic resource allocation. Rather, the model automatically classifies workloads and manages resources via real-time learning (RL) and adjustment. Results: The ART-driven model improved workload categorization, resource allocation and system throughput for all experiments, with particularly good adaptive properties in the face of changes in cloud conditions. Conclusions: The proposed ART-based approach to autonomic cloud systems enhances real- time, adaptive workload management; improves resource utilization and scalability; as well mitigates dynamicworkload concerns automatically without human intervention.