Advanced communication technologies (e.g., 5G) probably elicit a complete change of network and its applications. For example, a growing number of services begin shifting from central clouds to vast mobile devices, as the hybrid use of cloud and fog computing technologies can provide enhanced quality of service and efficient utilization of resources. However, to design such complex hybrid cloud-fog (HCF) systems, it remains a challenge to implement time-consuming modeling and inefficient evaluation in its early design stage based on the conventional simulation or practical experimentation. Therefore, how to reduce design cost and improve development efficiency becomes a crucial issue in the process of designing large-scale HCF systems. To address the issue, this paper proposes a novel modeling framework for large-scale HCF systems based on a high-level formal language, i.e. performance evaluation process algebra (PEPA). Toward the key components of an HCF system, the proposed framework includes three crucial model prototypes: compositional architecture model , abstract communication model and scheduling model . Moreover, the scheduling model is designed with a novel smart scheduling scheme that integrates two atomic scheduling algorithms and a decision module to make an efficient algorithm selection. The smart scheduling algorithm can well adapt the HCF systems by yielding stable and fast response to end-users, particularly under dynamical system conditions. Finally, the framework is the first research achieving the full potential of formal methods to implement industry-level modeling and evaluation.

中文翻译：

---

混合云雾系统的高效形式化建模框架

先进的通信技术（例如5G）可能会引起网络及其应用的彻底改变。例如，越来越多的服务开始从中央云转移到庞大的移动设备，因为云和雾计算技术的混合使用可以提供更高的服务质量和资源的有效利用。但是，要设计这种复杂的混合云雾（HCF）系统，在常规设计或实际实验的基础上，要在其早期设计阶段实现耗时的建模和低效评估仍然是一个挑战。因此，如何降低设计成本，提高开发效率成为设计大型HCF系统过程中的关键问题。为了解决这个问题，本文提出了一种基于高级形式语言即性能评估过程代数（PEPA）的大规模HCF系统的新型建模框架。对于HCF系统的关键组件，建议的框架包括三个关键的模型原型：组成架构模型 ， 抽象沟通模型 和 调度模型 。此外，利用新颖的智能调度方案设计了调度模型，该方案集成了两种原子调度算法和决策模块，可以进行高效的算法选择。通过对最终用户产生稳定而快速的响应，尤其是在动态系统条件下，智能调度算法可以很好地适应HCF系统。最后，该框架是第一个充分发挥形式方法实现行业级建模和评估潜力的研究。