Design methodologies and algorithms for survivable C-RAN

Abstract

In centralized/cloud radio access networks (C-RANs), baseband units (BBUs) are decoupled from remote radio units (RRUs) and placed in BBU hotels. In this way baseband processing resources can be shared among RRUs, providing opportunities for radio coordination and cost/energy savings. However, the failure of a BBU hotel can affect a large number of RRUs creating severe outages in the radio segment. For this reason, the design of a resilient C-RAN is imperative. In this paper, an extension of the facility location problem (FLP) is proposed to find the placement of BBU hotels that guarantees survivability against single hotel failure while the delay is minimized. Different strategies are proposed based on heuristic and integer linear programming (ILP) to solve the survivable BBU location problem and optimizing the sharing of backup resources. The results compare the proposed methodologies in terms of the costs of the BBU placement by referring to different network topologies. The heuristic algorithm is shown to find solutions close to those obtained by the ILP, although evidencing different contributions that are suitably discussed.