Towards an efficient, fair, and QoS-aware 5G network infrastructure

5G networks promise to deliver high link capacity, ultra-low latency, massive connectivity, and high availability. Compared to 4G, 5G integrates multiple advanced technologies, including network slicing, small cell deployment, beam forming, and software-defined RAN, to enhance performance. The 5G infrastructure consists of the Radio Access Network(RAN) and the Core Network. In particular, 5G RAN is usually the bottleneck resource and determines the user’s link capacity in the last mile. However, existing RAN resource management approaches cannot be directly applied to 5G networks with these new technologies, and often results in suboptimal performance. In this dissertation, I present three systems that address the key challenges in modern 5G networks in terms of load balancing, resource allocation, and congestion control. In the introductory chapter, I will discuss the background in 5G network architecture and 5G applications. In the second chapter, I will present RadioWeaver, a load-balancing system in the multi-cell deployment with 5G slicing. RadioWeaver defines a novel load balancing criteria across slices to identify which cells are overloaded despite the fact that different slices optimize for different criteria. By co-designing load balancing with dynamic quota allocation, RadioWeaver improves the optimization objective of every slice. In the third chapter, I will present RadioSaber, a practical and efficient channel-aware 5G RAN slicing system. RadioSaber aims to schedule wireless resources in a channel-aware manner at both the inter-slice scheduler and the enterprise scheduler. Furthermore, it allows every slice to customize its scheduling policy to optimize its own objective. In the fourth chapter, I will present Octopus, a congestion control protocol for real-time applications, which achieves both high throughput and low latency. Octopus leverages in-network content adaptation to reduce the data rate of multimedia streams when the link capacity drops abruptly, so as to maintain low queuing delay.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Electrical & Computer Engr

Grantor

University of Illinois Urbana-Champaign

Year dc:date

2025

Author dc:creator

• Chen, Yongzhou

Contributors dc:contributor

• Mittal, Radhika
• Godfrey, Philip Brighten
• Choudhury, Romit Roy
• Srikant, Rayadurgam

• 5G networks
• cellular networks
• mobile networks
• edge computing
• resource allocation
• load balancer
• congestion control
• video streaming
• machine learning

Statement dc:rights

• Copyright 2025 Yongzhou Chen

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/129395

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F129395