Optimal Collaborative Configuration Strategy of IaaS Resources Under Multiple Pricing Models for Maximizing SaaS Providers’ Expected Revenue

Current cloud resource configuration schemes at the Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service (PaaS) levels often result in frequent Quality of Service (QoS) violations, low resource utilization, and inadequate revenue assurance for Software-as-a-Service (SaaS) providers. To overcome these limitations, this paper proposes a novel two-stage, optimal collaborative configuration strategy for IaaS resources, designed explicitly to maximize SaaS providers’ expected revenue under three prevalent IaaS pricing models. In the first stage, each SaaS provider determines its initial optimal resource demand using historical user data. In the second stage, resources are dynamically reallocated collaboratively among SaaS providers experiencing resource surpluses and deficits. This strategy achieves a dual objective: maximizing the SaaS provider’s expected revenue while enabling the IaaS provider to enhance utilization through more precise resource allocation—all while ensuring zero QoS violations at the IaaS provider level and a drastically reduced probability of SaaS-to-user QoS violations. We instantiate this framework by deriving optimal collaborative configuration strategies for three prevalent IaaS pricing models: Fixed-price (OCCS_FI), Segmented-price (OCCS_SI), and Dynamic-price (OCCS_DI). Theoretical analysis and comprehensive experimental evaluations confirm the efficacy of our proposed strategies. Under conditions of stochastic user demand, our strategies ensure no QoS violations are triggered at the IaaS provider level, while seeking to maximize the expected revenue for SaaS providers and maintain high resource utilization. This is achieved by determining an optimal initial resource purchase that accounts for demand uncertainty, followed by a collaborative reallocation mechanism that mitigates shortages. Combined, these measures reduce the probability and impact of SaaS-to-user QoS violations to a negligible level.