In this paper, we jointly consider cloud radio access networks (C-RAN) and mobile cloud computing (MCC) in a holistic framework. In particular, we study how to dynamically operate C-RAN to enhance the end-to-end performance of MCC services in next generation wireless networks. An intrinsic challenge in such system is that channel state information (CSI) is outdated. With delayed CSI, only sub-optimal C-RAN operations can be made if deterministic optimization techniques are applied directly. We formulate the topology configuration and rate allocation problem with delayed CSI under a stochastic optimization framework.

Such framework maximizes MCC services’ sum throughput with constraints on the response latency experienced by each MCC user. We propose an optimal policy for the stochastic optimization problem, which has the merit of low computation cost. Offline and online algorithms are developed based on the optimal policy. Using simulation results, we show that, with the emergence of MCC and C-RAN technologies, the design and operation of future mobile wirelessnetworks can be significantly affected by cloud computing, and the proposed scheme is capable of achieving substantial performance gains over existing schemes.