Efficient and Scalable Design of Resource Allocation and Orchestration Solutions Towards Building Beyond 5G Networks

5G网络代表了通信网络设计的突破，因为它们能够提供单一平台，实现各种不同的服务，从增强型移动宽带通信到扩展现实、触觉、自动驾驶和物联网。然而，目睹对新服务的需求不断增长，并预测未来十年内新技术的发展，我们已经有可能预见到超越5G （B5G）的需求，并设计一种集成尖端技术的新型架构，以满足个人和社会层面的新需求。B5G将提供信息和通信技术基础设施，使最终用户能够感知自己被“巨大的人工大脑”包围，提供虚拟零延迟服务、无限存储和巨大的认知能力。B5G网络支持的业务具有非常多样化的关键性能指标，因此设计一个单一平台以有效的方式支持所有这些业务是一项非常具有挑战性的任务。此外，由于B5G支持的大多数服务都是通过运行在异构分布式基础设施上的完全虚拟化功能来启用的，因此它们都需要由一个通用的编排框架来管理。然而，与B5G相关的大量开放研究问题和挑战可以确定，包括但不限于整体通信、计算和缓存虚拟资源分配、端到端服务供应、多域编排问题、基础设施提供商的利润最大化、延迟关键服务的支持。因此，我的发现研究计划将尝试阐述和解决与B5G网络采用的资源分配和编排解决方案的设计和运营相关的关键研究挑战和问题。通过联合通信、计算和缓存虚拟资源分配以及智能和经济高效的资源编排，在B5G中探索创新的资源配置解决方案，将实现这一长期目标。资源分配和编排将首先通过理论方法解决，包括数学优化，以开发和分析最佳解决方案和性能界限。接下来，将使用启发式、基于人工智能的集中式或分布式方法设计和验证不太复杂的次优和自适应解决方案。最后，将建立定制的测试平台，以在现实环境中演示所提出的解决方案。由于该研究计划涉及对几名研究生的培训，因此预计该研究计划将在B5G网络的学术知识和工业进步方面带来好处。研究成果将通过开放获取预印本共享，由此产生的知识产权（和专利）将转移到加拿大工业。

5G networks represent a breakthrough in the design of communication networks, for their ability to provide a single platform enabling a variety of different services, from enhanced mobile broadband communications to eXtended reality, haptic, automated driving and Internet-of-Things. Yet, witnessing the ever increasing demands for new services and predicting the evolution of new technologies within a decade from now, it is already possible to envision the need to move beyond 5G (B5G) and design a novel architecture integrating cutting-edge technologies to satisfy new needs at both individual and societal level. B5G will provide an Information and Communication Technology infrastructure that will enable the end users to perceive themselves as surrounded by a "huge artificial brain" providing virtual zero-latency services, unlimited storage, and immense cognition capabilities. The services enabled by B5G networks are characterized by very diverse sets of key performance indicators so that the design of a single platform enabling all of them in an efficient manner is a very challenging task. Besides, since most services enabled by B5G are enabled through fully virtualized functions that run overs heterogeneous distributed infrastructures, they all need to be managed by a common orchestration framework. However, a plethora of open research issues and challenges related to B5G can be identified, including but not limited to, holistic communication, computing and caching virtual resource allocation, end-to-end service provisioning, multi-domain orchestration issues, profit maximization for infrastructure providers, support of latency critical services. Therefore, my discovery research program will attempt to elaborate and address key research challenges and problems related to the design and operation of resource allocation and orchestration solutions to be adopted by B5G networks. This long-term objective will be attained via exploring innovative resource provisioning solutions in B5G through joint communication, computing and caching virtual resource allocation as well as intelligent and economically efficient resource orchestration. Resource allocation and orchestration will be first tackled through theoretical approaches including mathematical optimization to develop and analyze optimal solutions and performance bounds. Following, less-complex suboptimal and adaptive solutions using heuristic, AI-based centralized or distributed approaches will be designed and validated by simulation tools. Finally, customized testbeds will be built to demonstrate the proposed solutions in realistic environments. The benefits of this research program are expected in both academic knowledge and industrial advances in B5G networks as it involves the training of several graduate students. The research results will be shared via open access preprints and the resulting intellectual property (and patents) will be transferred to Canadian industries.