| YANG Xining (杨喜宁)**,ZHOU Yiqing* **,CHEN Yang**.[J].高技术通讯(英文),2022,28(4):384~391 |
|
| An efficient cross-layer buffer management optimization scheme for 5G protocol stack |
| |
| DOI：10.3772/j.issn.1006-6748.2022.04.006 |
| 中文关键词: |
| 英文关键词: cross-layer optimization, 5G protocol stack, evaluation framework, OpenAirInterface (OAI) |
| 基金项目: |
| Author Name | Affiliation | | YANG Xining (杨喜宁)** | (*Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100083, P.R.China)
(**School of Computer Science and Technology, University of Chinese Academy of Science, Beijing 100190, P.R.China) | | ZHOU Yiqing* ** | (*Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100083, P.R.China)
(**School of Computer Science and Technology, University of Chinese Academy of Science, Beijing 100190, P.R.China) | | CHEN Yang** | (*Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100083, P.R.China)
(**School of Computer Science and Technology, University of Chinese Academy of Science, Beijing 100190, P.R.China) |
|
| Hits: 609 |
| Download times: 614 |
| 中文摘要: |
| |
| 英文摘要: |
| It is extremely challenging for the 5G User Equipment (UE) to meet the requirement of low-latency data transmission with higher achievable data rates. And user plane processing of 5G protocol stack (PS) is one of the dominating components for end-to-end data transmission in the network system. In this paper, a cross-layer buffer management scheme (CLBM) is proposed. CLBM adopts a zero-copy technique for protocol data unit (PDU) processing between protocol layers and allows to improve the memory operation efficiency significantly with reduced processing latency and CPU usage. Moreover, the PS performance profiling (PSperf) tool, a general evaluation framework for the performance measurement and analysis of PS, is implemented based on the OpenAirInterface (OAI) 5G platform. The evaluation result shows that compared with the PS of OAI the CLBM strategy reduces the CPU usage of RLC, PDCP, and MAC layer processing significantly up to 20.6%, 63.4%, and 38.8%, respectively. In result, the processing delay of the whole user plane of PS also has been reduced distinctly at various offered traffic load. |
|
View Full Text
View/Add Comment Download reader |
| Close |
|
|
|
