Background and Motivation
Modern network infrastructures are increasingly distributed, combining centralized cloud environments with compute resources deployed closer to the network edge. This architectural shift is driven by the growing demand for latency-sensitive and bandwidth-intensive applications such as immersive media, connected vehicles, intelligent sensing systems, and AI-driven services. To support these applications, communication networks are evolving toward highly flexible and programmable infrastructures where networking and compute resources can be dynamically orchestrated across heterogeneous environments.
A critical component of these systems is the data plane, which handles the actual forwarding and processing of user traffic. In modern cellular and edge-cloud deployments, the data plane can be distributed across multiple locations in the network, enabling traffic processing closer to users in order to reduce latency, improve scalability, and optimize resource utilization. However, orchestrating distributed data plane functions across geographically dispersed infrastructure introduces significant challenges.
Current open-source platforms and emerging standards provide mechanisms to deploy and manage network functions across cloud and edge environments. While these solutions enable basic deployment and scaling, they often rely on centralized orchestration logic and static placement strategies that do not fully exploit the potential of distributed infrastructures. As a result, limitations may arise in terms of scalability, responsiveness to changing network conditions, and the ability to maintain high Quality of Experience (QoE) for end users.
This thesis aims to investigate the orchestration of distributed data plane functions for user traffic in edge-cloud network deployments. The objective is to analyze the limitations of current solutions, explore how distributed orchestration mechanisms can be improved, and propose enhancements that enable more efficient resource utilization and improved user experience. The work will be conducted using existing open-source platforms and realistic experimental environments, with a focus on practical implementation and evaluation.
Expected Outcomes
The student will work to:
- Conduct a comprehensive literature review on distributed data plane architectures, edge-cloud networking, and orchestration frameworks, identifying current limitations in existing solutions and standards;
- Analyze existing open-source implementations of data plane deployment and orchestration mechanisms, studying how user traffic processing functions are currently instantiated, placed, and managed across distributed infrastructures;
- Deploy and experiment with a reference platform in a controlled environment to understand the operational behavior and performance characteristics of current approaches;
- Identify specific limitations related to orchestration, scalability, adaptability, or Quality of Experience in distributed deployments;
- Design and implement an improvement to one aspect of the orchestration framework (e.g., smarter function placement, distributed control logic, adaptive scaling, or QoE-aware resource management);
- Evaluate the proposed solution through experimentation in a realistic test environment, measuring performance improvements such as latency, scalability, resource efficiency, or user experience metrics.
Requirements
- Familiarity with networking fundamentals. Knowledge of 5G/6G concepts is a plus;
- Experience with Linux environments and system-level tools;
- Knowledge of containerization (e.g., Docker, Kubernetes) is a plus;
- Background or strong interest in machine learning, optimization, or control systems;
- Programming experience (e.g., Python, C++, or similar languages) for prototyping and system integration;
- Willingness to work with real-world or realistic testbed environments and to engage with ongoing research projects at SPEAR Lab.
Related Work
Goshi, E., Harkous, H., Ahvar, S., Pries, R., Mehmeti, F., and Kellerer, W. 2024. Joint UPF and Edge Applications Placement and Routing in 5G & Beyond. IEEE MASS Workshop on Mobile Access, Sharing & Services (MASS '24). IEEE. https://arxiv.org/abs/2407.07669
Leyva-Pupo, I., Cervelló-Pastor, C., Anagnostopoulos, C., and Pezaros, D. P. 2022. Dynamic UPF Placement and Chaining Reconfiguration in 5G Networks. Computer Networks, vol. 215. Elsevier. DOI: 10.1016/j.comnet.2022.109200
Alevizaki, V.-M., Anastasopoulos, M., Tzanakaki, A., and Simeonidou, D. 2021. Dynamic Selection of User Plane Function in 5G Environments. In Proceedings of the 2021 International Conference on Optical Network Design and Modeling (ONDM '21). IFIP, Athens, Greece.
Bartolomeo, G., Mohan, N., et al. 2023. Oakestra: A Lightweight Hierarchical Orchestration Framework for Edge Computing. In Proceedings of the 2023 USENIX Annual Technical Conference (USENIX ATC '23). USENIX Association. https://www.usenix.org/conference/atc23/presentation/bartolomeo
Pham, Q.-V., Fang, F., Ha, V. N., et al. 2020. A Survey of Multi-Access Edge Computing in 5G and Beyond: Fundamentals, Technology Integration, and State-of-the-Art. IEEE Access, vol. 8, 116974–117017. DOI: 10.1109/ACCESS.2020.3001277
