A Multi-Criteria Framework for Performance, Security, And Cost Optimization in Cloud Computing Applications

Abstract

Cloud computing has become a major computing paradigm for delivering scalable infrastructure, application platforms, and software services. However, selecting and managing cloud resources remains challenging because application performance, security requirements, and operational expenditure frequently compete with one another. Configurations that provide low response times may increase resource costs, while aggressive cost reduction may introduce performance degradation or unacceptable security risks. This paper proposes a Multi-Criteria Cloud Application Optimization Framework (MCCAF) for evaluating cloud application configurations through integrated performance, security, and cost assessment. The framework combines workload characterization, security-policy enforcement, resource configuration, multi-objective optimization, and sensitivity analysis. Its mathematical formulation minimizes application response time and total operating cost while maximizing verified security-control coverage, subject to application-specific service-level requirements. A Pareto-based decision process is proposed to identify non-dominated configurations before selecting an appropriate solution according to stakeholder priorities. The framework also specifies an experimental methodology using CloudSim or a comparable test environment, benchmark scheduling methods, repeated workload trials, and statistical evaluation. Unlike a single-score optimization approach, the framework treats mandatory security requirements as constraints and evaluates the trade-offs between competing objectives. The proposed methodology establishes a foundation for reproducible cloud application optimization, although its effectiveness and superiority require empirical validation. The study contributes an application-oriented decision framework for researchers and cloud administrators seeking balanced performance, security, and economic efficiency.