Joint optimization of computing and radio resource under outage QoS constraint in C-RAN

Abstract

Cloud Radio Access Network (C-RAN) enhances the utilization of computing and radio resources by concentrating Baseband Unit (BBU) at a data center. It is essential to find a flexible way to manage the computing resource since processing power has become the majority of system power in C-RAN. The computing resource demand of users can be scaled according to the modulation bits and other parameters in transmission. To this end, the paper aims at minimizing the system power consumption by jointly allocating the computing and radio resource under the outage Quality of Service (QoS) constraint. Using outage-QoS allows the resource allocation need not update frequently to keep up with the unstable channel state. The power minimization problem is formulated as a Mixed Integer Non-linear Problem (MINLP). To solve the complex MINLP, we first fix the virtual machine-user mapping relationship, then transform the MINLP into a Geometric Programming (GP) by omitting the Remote Radio Heads (RRHs) power budget. Finally, we propose an over-budget RRH reallocation greedy algorithm to reallocate resource for the users served by the heaviest over-budget RRH. Simulation results show the advantage of our scheme on system power saving and efficiency on resource management.