A survey on state-of-the-art experimental simulations for privacy-preserving federated learning in intelligent networking

Abstract

<abstract> <p>Federated learning (FL) provides a collaborative framework that enables intelligent networking devices to train a shared model without the need to share local data. FL has been applied in communication networks, which offers the dual advantage of preserving user privacy and reducing communication overhead. Networking systems and FL are highly complementary. Networking environments provide critical support for data acquisition, edge computing capabilities, round communication/connectivity, and scalable topologies. In turn, FL can leverage capabilities to achieve learning adaptation, low-latency operation, edge intelligence, personalization, and, notably, privacy preservation. In our review, we gather relevant literature and open-source platforms that point out the feasibility of conducting experiments at the confluence of FL and intelligent networking. Our review is structured around key sections, including the introduction of FL concepts, the background of FL applied in networking, and experimental simulations covering networking for FL and FL for networking. Additionally, we delved into case studies showcasing FL potential in optimizing state-of-the-art network optimization objectives, such as learning performance, quality of service, energy, and cost. We also addressed the challenges and outlined future research directions that provide valuable guidance to researchers and practitioners in this trending field.</p> </abstract>

A Survey of Intelligent End-to-End Networking Solutions: Integrating Graph Neural Networks and Deep Reinforcement Learning Approaches

“…Therefore, before focusing on other potential issues in E2E networkin one key research is the selection of optimization algorithms that handle complex grap structured topologies and extract data to support self-organizing capabilities [14,15]. Previous works supported by standardization, academia, and industry experts, a coming to conduct the creation of cutting-edge testbeds and simulation tools for netwo intelligence [16][17][18][19]. The motivation from existing testbeds has guided researchers t wards integrating three key objectives, namely zero-touch autonomy, topology-awa scalability, and long-term efficiency, into network and service management [20,21].…”

“…Previous works supported by standardization, academia, and industry experts, are coming to conduct the creation of cutting-edge testbeds and simulation tools for network intelligence [16][17][18][19]. The motivation from existing testbeds has guided researchers towards integrating three key objectives, namely zero-touch autonomy, topology-aware scalability, and long-term efficiency, into network and service management [20,21].…”