Breaking Platform Silos: Dynamic Multi-cloud Cooperation with Long-term Profit and Fairness Enhancement

Abstract

In the era of geographically distributed workloads and diverse computational demands, ranging from latency-sensitive scientific simulations to large-scale language model training, individual cloud providers often face limitations in geographic reach and incomplete coverage of heterogeneous resources. This platform silo effect constrains their ability to deliver optimal service-level agreements (SLAs) and achieve sustainable profits. Existing multi-cloud frameworks are predominantly designed with master–slave control architectures, which hinder the dynamic participation of providers. Furthermore, most current scheduling strategies overlook both the future profit potential arising from resource heterogeneity and the need for fair profit-sharing mechanisms. We address these challenges with a Dynamic Multi-cloud Cooperation framework that breaks platform silos through three tightly coupled stages: (i) Pre-cooperation: a peer-to-peer collaboration framework (PCM) with an adaptor component to unify heterogeneous interfaces, enabling low-cost, on-demand participation of diverse providers; (ii) In-cooperation: a long-term profit–aware scheduling method (DMCLP) that jointly optimizes current profits and predicted future profits by maintaining high resource availability; (iii) Post-cooperation: a Shapley-value–based profit allocation scheme ensuring that each provider receives a fair share proportional to its true marginal contribution. Extensive simulations across diverse task types, provider configurations, and workload scales demonstrate that DMCLP increases provider profits by an average of 5.33%, with improvements reaching up to 40.33% in extreme cases, while simultaneously improving resource high availability by an average of 3.4% and user satisfaction by 6.11% compared to baselines. These results confirm the framework’s effectiveness in fostering sustainable multi-cloud cooperation.

Keywords

JointCloud computing / peer collaboration / resource scheduling / long-term profit / fairness

RIGHTS & PERMISSIONS

Higher Education Press 2026

Just Accepted

This article has successfully passed peer review and final editorial review, and will soon enter typesetting, proofreading and other publishing processes. The currently displayed version is the accepted final manuscript. The officially published version will be updated with format, DOI and citation information upon launch. We recommend that you pay attention to subsequent journal notifications and preferentially cite the officially published version. Thank you for your support and cooperation.