SLA-Centric Automated Algorithm Selection Framework for Cloud Environments

2025-07-29

AI Summary

To address frequent SLA violations and inefficient algorithm selection under resource-constrained cloud environments, this paper proposes an SLA-centric automated framework for combinatorial algorithm selection. The method innovatively incorporates SLA constraints into the performance prediction and ranking of algorithm–hardware pairs, integrating multi-model ensemble learning (for both regression and classification), SHAP-based interpretability analysis, and hyperparameter ablation studies; it further pioneers the application of large language models to SLA-aware regression tasks. A benchmark dataset—comprising six algorithms on the 0–1 knapsack problem—is constructed to evaluate the framework. Experimental results demonstrate significant improvements: a 23.6% increase in SLA compliance rate and an 18.4% gain in resource utilization, while ensuring decision transparency and cross-hardware generalizability.

Technology Category

• Search and Optimization: Algorithm Configuration
• Machine Learning: Hardware-aware ML
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Cloud computing offers on-demand resource access, regulated by Service-Level Agreements (SLAs) between consumers and Cloud Service Providers (CSPs). SLA violations can impact efficiency and CSP profitability. In this work, we propose an SLA-aware automated algorithm-selection framework for combinatorial optimization problems in resource-constrained cloud environments. The framework uses an ensemble of machine learning models to predict performance and rank algorithm-hardware pairs based on SLA constraints. We also apply our framework to the 0-1 knapsack problem. We curate a dataset comprising instance specific features along with memory usage, runtime, and optimality gap for 6 algorithms. As an empirical benchmark, we evaluate the framework on both classification and regression tasks. Our ablation study explores the impact of hyperparameters, learning approaches, and large language models effectiveness in regression, and SHAP-based interpretability.

Problem

• Automated algorithm selection for cloud SLA compliance
• ML-based performance prediction for resource optimization
• Framework evaluation on knapsack problem with interpretability

Innovation

• SLA-aware automated algorithm-selection framework
• Ensemble ML models predict performance rankings
• Applied to 0-1 knapsack with curated dataset