The cloud radio access network (CRAN) architecture provides a pool of shareable centralized baseband processing unit (BBU), which is responsible for processing the entire network load exerted by remote radio heads (RRHs), regarded as ‘RRH- requests’. In this dissertation, we consider each BBU has only one core-computing unit (CCU) that is fitted with a single core general-purpose processor (GPP). Computing resource (CR) allocation has been widely investigated in CRAN, which decides the system’s resource and energy efficiency. However, these schemes allo- cate CR without considering GPP operating temperature and thus are not feasible in practice. This is because none GPP can process any task once its operating temperature reaches the permissible limit. Therefore, it is necessary to optimize the CR utilization while taking the GPP operating temperature into considera- tion. On the other hand, to optimize CR utilization, less number of BBUs has to execute the same amount of RRH-requests. This causes dramatic rise in GPP’s op- erating temperature and imposes computational power (CP) efficiency problem.
Sumarga Kumar Sah Tyagi
Sumarga Kumar Sah Tyagi received PhD. from Institute of Computing Technology, Chinese Academy of Sciences. Currently, he is with the School of Electronic and Information Engineering, Zhongyuan University of Technology, Zhengzhou, China, working as a Lecturer. His research interests are CRAN, Machine learning, DNN techniques for aerospace & cloud.
Number of Pages:
LAP LAMBERT Academic Publishing
Cloud RAN, BBU pooling, computing resource allocation, compu- tational power efficiency, Energy-efficiency, RRH-to-BBU mapping, performance- power tradeoff, Workload Balancing, thermal balancing, DVFS.
COMPUTERS / Networking / General