基於協同傳輸的省電排程技術於C-RAN網路

Energy-Efficient Downlink Scheduling with Joint Processing in Cloud Radio Access Networks

許經國(Ching-Kuo Hsu)

指導教授 ： 曾煜棋 ； 梁家銘

國立交通大學/資訊學院/資訊科學與工程研究所/博士(2020年)

摘要

在5G通訊中，C-RAN被用來提供寬頻服務。其將計算單元BBU(Baseband Units)從基地台(base stations)中分離並置於集中管理的雲。近期的研究探討利用聯合處理(joint processing)技術讓行動裝置(user equipments,UEs)從多個協同傳輸的基地台接收資源。協同傳輸可能更耗電。雖然3GPP提出非連續接收機制(DRX)調和行動裝置週期性關閉無線介面，如何妥善處理C-RAN下的聯合處理與DRX機制仍是一個開放議題。另一方面，隨著行動裝置增加，網路負載不均的狀況會越來越常發生。因此，資料卸載是在這樣的情境下有效平衡基地台間負載的方法。而且透過C-RAN，資料卸載會更便利。當行動裝置在熱點且因為網路壅塞無法被服務時，C-RAN可以將他們的資料請求卸載到鄰近基地台以平衡基地台間的負載。因此，我們針對C-RAN提出資源分配與省電排程機制，其包含三項研究主題，第一個研究主題主要針對動態擇點協作方式。第二個研究主題主要針對資料卸載與動態擇點協作方式。第三個研究主題主要針對聯合傳輸方式。
在第一個研究主題中，我們首先說明此問題為NP-complete並提出兩個有效率的動態擇點協作方法，分別是服務比例法與排程花費法。服務比例法基於服務時間比例確保行動裝置連續接收資源, 尤其是落在基地台交界的裝置。而排程花費法藉由兩個排程花費評效來有效平衡節能效率與系統產出。
在第二個研究主題中，我們提出一個基於資料卸載的動態擇點協作排程方法，其利用優化能耗花費來排程。這個方法有三個階段。第一階段決定DRX週期長度以避免資源競爭。第二階段根據基地台的卸載潛力執行資料卸載。第三階段使用三個特別的規則來妥善決定DRX參數。
在第三個研究主題中，我們提出一個節能聯合傳輸排程方法，其妥善配對行動裝置與基地台，並考慮傳輸品質來改善排程。這個方法有三個階段。第一階段決定DRX週期長度以避免資源競爭。第二階段根據傳輸品質與傳輸速率決定排程。第三階段使用三個特別的規則來妥善決定DRX參數。

關鍵字

5G ； C-RAN ； 非連續接收機制 ； 動態擇點協作 ； 聯合傳輸 ； 資料卸載 ； 資源排程 ； 節能 ； 服務品質 ； 睡眠排程

並列摘要

In 5G mobile communications, Cloud radio access networks (C-RAN) is proposed to provide broadband services. It separates computation entities, i.e., baseband units (BBUs), from base stations (BSs) and puts all BBUs in a centralized cloud. Existing researches have investigated that how to enhance user equipments (UEs) to receive data from multiple collaborative cells by leveraging the joint processing (JP) technology. Collaborative transmission may cause higher energy consumption for UEs. Although 3GPP defines the discontinuous reception (DRX) mechanism by regulating UEs to close their radio interfaces periodically, how to well coordinate DRX with JP in C-RAN is still an open issue. On the other hand, when the number of UEs grows, network loads may become imbalanced more often. Thus, data offloading is an efficient way to balance cells' loads in such scenarios, and it becomes more convenient through C-RAN. When UEs are in a hotspot, they may not be served by the network because of congestion, C-RAN can offload their data requests to the neighboring cells to balance the cells' loads. Based on the above problems, we propose the DRX optimization problem under given QoS constraints of UEs in C-RAN with JP, which is composed of three works. The first work discusses the DRX optimization problem under given QoS constraints of UEs in a dynamic point selection (DPS) C-RAN. The second work discusses the DRX optimization problem under given QoS constraints of UEs in C-RAN with DPS and data offloading. The third work discusses the DRX optimization problem under given QoS constraints of UEs in a joint transmission (JT) C-RAN.
In the first work, we prove that this problem is NP-complete, and then propose two effective DPS solutions, serving-ratio (SR) scheme and cost-aware (CO) scheme. SR serves UEs according to 'serving ratio' to make UEs receive data continuously, especially for those in cell intersections. On the other hand, CO exploits two special scheduling cost metrics to balance energy and throughput efficiency.

In the second work, we propose an offloading-based DPS scheduling scheme by exploiting minimal energy cost. The scheme has three stages. The first stage determines the DRX cycle for avoiding resource competition. The second stage offloads data requests according to offloading potential of cells. The third stage uses three special rules to better determine the DRX parameters.

In the third work, we propose an energy-efficient JT scheduling scheme by better pairing UEs and cells with consideration of joint transmission quality. The scheme has three stages. The first stage determines the DRX cycle for avoiding resource competition. The second stage determines the data scheduling order according to UEs' MCS levels and achievable data rate. The third stage uses three special rules to better determine the DRX parameters.

並列關鍵字

5G ； Cloud Radio Access Network (C-RAN) ； discontinuous reception (DRX) ； dynamic point selection (DPS) ； joint transmission (JT) ； data offloading ； resource allocation ； power saving ； quality-of-services (QoS) ； sleep scheduling

參考文獻

[1] D. Sabella et al., "RAN as a service: Challenges of designing a flexible RAN architecture in a cloud-based heterogeneous mobile network," Future Network & Mobile Summit, pp. 1–8, 2013.

[2] S. S. Solimana and B. Songb, "Fifth generation (5G) cellular and the network for tomorrow: cognitive and cooperative approach for energy savings," Journal of Network and Computer Applications, vol. 85, no. 1, pp. 84–93, 2017.

[3] V. Fernandez-Lopez, B. Soret, K. I. Pedersen, J. Steiner, and P. Mogensen, "Sensitivity analysis of centralized dynamic cell selection," IEEE Vehicular Technology Conference (VTC), pp. 1–5, 2016.

[4] M. Eguizabal and A. Hernandez, "Joint dynamic resource allocation and load balancing-cell selection in LTE-A HetNet scenarios based on Type 1 inband relay deployments," Computer Networks, vol. 100, pp. 90–109, 2016.

[5] A. De Domenico, L. Diez, R. Aguero, D. Ktenas, and V. Savin, "EMF-aware cell selection in heterogeneous cellular networks," IEEE Communications Letters, vol. 19, no. 2, pp. 271–274, 2015.