Cost models for BaseBand Unit (BBU) hotelling: From local to cloud

In this work we model and compare three different strategies for BaseBand Unit (BBU) hotelling: BBU stacking, BBU pooling, and Cloud RAN. We propose the building-block models for every strategy and elaborate on a cost model in order to evaluate their benefits. We also propose shelf and rack design for the implementation of the presented models. The assessment includes cost comparisons and sensitivity analysis with respect to pooling and virtualization gains. We obtain that the best option in terms of cost savings is the cloud-based BBU hotel. However, we show that, under equal pooling gains, the benefits are mainly based on the implementation of the technology and the shelf/rack design.

I. Introduction

In recent years, there has been a rapid growth of mobile data traffic due to new bandwidth-intensive applications (as video streaming) and due to the increasing number of users [1]. Supporting such high traffic demand requires an increase in computational resources and transmission capacity. To face this challenge, mobile operators are investigating novel solutions for cost-effective network expansion and upgrade. Consolidation and centralization of network equipment and functions have become key enablers to this end. Thanks to centralization, collateral systems, i.e., cabinets, racks, power supplying (voltage transformers, rectifiers, backup batteries), cooling (ventilation or air conditioning), backplanes, and aggregation gateways, can be re-implemented to save in energy and costs. New technologies such as cloud computing and virtualization provide an effective way to share the processing resources.