近幾年來，5G技術和物聯網發展密不可分，其中網絡功能虛擬化是一項重要的發展。當中功能拆分這項方法可以將傳統的基地台功能虛擬化並集中部分功能至控制器的中央單元上，其餘功能則委託給更靠近天線的分布式單元，這項方法非常適合用於大規模部屬的物聯網設備。
我們提出了研究針對5G功能拆分的研究。為簡化起見，我們假設系統擁有二個服務速率不為零的拆分與一個服務速率為零的預備拆分。如果要在服務速率不為零的二個拆分之間切換，來源拆分必須先切換至預備拆分，然後再自預備拆分切換至目的拆分。整個系統具有一個有限的佇列。每一拆分的停留時間皆是指數型分布。我們將封包分為兩種優先權並考慮兩種情境：(1) 具有延遲優先權的功能拆分，其中實施延遲優先權機制，亦即進入系統後，只要封包尚未被服務時，高優先權封包可以排在低優先權封包的前方，與(2)具有延遲與捨棄優先權的功能拆分，其中除了延遲優先權外，也實施基於預留的捨棄優先權機制，亦即系統會預留一些位置僅供高優先權的封包使用。
在這項研究中，我們首先運用四維馬可夫鏈來塑模所考慮的系統。第二，我們推導出模型的狀態平衡方程式，並使用迭代演算法來求得穩態機率分布。第三，我們使用C語言針對模型撰寫電腦模擬程式。第四，我們探討了不同參數對於系統的影響，最終在所有的研究案例當中，解析結果和電腦模擬的結果相當接近。

In recent years, 5G technology and the development of the Internet of Things (IoT) are inseparable, and Network Function Virtualization (NFV) is an important development. Functional split is an approach that virtualizes traditional base station functions and centralizes some functions on the Central Unit of the controller, while delegating the remaining functions to Distributed Unit closer to the antenna. This approach is ideal for IoT devices deployed at a large scale.
We present studies targeting functional split in 5G. For simplicity, we assume that there are two splits with non-zero service rate and one stand-by split with zero service rate. To switch among splits with non-zero service rate, the system must first switch from the source split to the stand-by split and then switch from the stand-by split to the destination split. This system is assumed to have a finite queue. The duration in each split is exponentially distributed. We classify packets into two priorities and consider two scenarios: (1) functional split with delay priority, where the delay scheme is enforced, i.e., after entering the system, as long as packets have not been serviced, high-priority packets are queued in front of low priority packets, and (2) functional split with delay priority and loss priority, where in addition to the delay priority, the loss priority scheme based on reservation is enforced, i.e., some places are reserved exclusively for high-priority packets.
In this study, first, we use the 4-dimensional Markov chain to model the system considered. Second, we derive the state balance equations of the model, and use the iterative algorithm to find the steady state probability distribution. Third, use C language to write simulation programs for the model. Fourth, we explore the effect of different parameters on the system. Finally, in all of the studied cases, the analytical results are close to the simulation results.

[1] T. Zhang, H. Qiu, L. Linguaglossa, W. Cerroni and P. Giaccone, "NFV Platforms: Taxonomy, Design Choices and Future Challenges," IEEE Transactions on Network and Service Management, vol. 22, no. 2, pp. 30 - 48, March 2021.
[2] M. Peng, C. Wang, V. Lau and H. V. Poor, "Fronthaul-Constrained Cloud Radio Access Networks: Insights and Challenges," IEEE Wireless Communications, vol. 22, no. 2, pp. 152 - 160, April 2015.
[3] C.-L. I, S. Kuklinskí and T. Chen, "A Perspective of O-RAN Integration with MEC, SON, and Network Slicing in the 5G Era," IEEE Network, vol. 34, no. 6, pp. 3 - 4, November/December 2020.
[4] L. Diez, A. M. Alba, W. Kellerer and R. Agüero, "A Quasi-Birth-Death Model for Functional Split in 5G Controllers," ICC 2021 - IEEE International Conference on Communications, vol. 9, pp. 14-23, August 2021.
[5] Z. Becvar, P. Mach, M. Elfiky and M. Sakamoto, "Hierarchical Scheduling for Suppression of Fronthaul Delay in C-RAN with Dynamic Functional Split," IEEE Communications Magazine, vol. 59, no. 4, pp. 95 - 101, April 2021.
[6] L. Diez, A. M. Alba, W. Kellerer and R. Agüero, "Flexible Functional Split and Fronthaul Delay: A Queuing-Based Model," IEEE Access, vol. 9, pp. 151049 - 151066, October 2021.
[7] G. Souza, A. Duarte, G. Moreira and F. Cruz, "Post-processing Improvements in Multi-objective Optimization of General Single-server Finite Queueing Networks," IEEE Latin America Transactions, vol. 21, no. 3, pp. 381 - 388, March 2023.
[8] J. X. Liao and X. W. Wu, "Resource Allocation and Task Scheduling Scheme in Priority-Based Hierarchical Edge Computing System," 2020 19th International Symposium on Distributed Computing and Applications for Business Engineering and Science (DCABES), pp. 16-19, 2020.
[9] H. Xu, X. Liu, W. G. Hatcher, G. Xu, W. Liao and W. Yu, "Priority-Aware Reinforcement-Learning-Based Integrated Design of Networking and Control for Industrial Internet of Things," IEEE Internet of Things Journal , vol. 8, no. 6, pp. 4668 - 4680, 15 March 2021.
[10] S. R. Mallick, V. Goswami, R. N. Dash, R. K. Lenka, S. Sharma and R. K. Barik, "A priority-reservation queueing-based approach for Blockchain-assisted smart-grid system," 2023 International Conference on Power Electronics and Energy (ICPEE), pp. 03-05, January 2023.
[11] Y. Arii, K. Yamaoka and K.-I. Baba, "A Two-stage Trunk Reservation Control Method corresponding to General Call Priority using Switch Outside Disaster Area," 2023 International Conference on Computing, Networking and Communications (ICNC), pp. 20-22, February 2023.
[12] A. Grigorjew, F. Metzger, T. Hoßfeld, J. Specht, F.-J. Götz, F. Chen and J. Schmitt, "Bounded Latency with Bridge-Local Stream Reservation and Strict Priority Queuing," 2020 11th International Conference on Network of the Future (NoF), pp. 12-14, October 2020.