The evolution of the wireless networks to a heterogeneous architecture is paralleled with an evolution in the design of the base station. In one evolutionary trend, compact base stations that combine baseband processing and radio function in the same enclosure became available on the market and are now widely known as 'small cells'. In a second evolutionary trend, macro cell base stations adopted a split-architecture that separate the baseband from the radio with a high-speed interface. The most common interface protocol between the baseband and remote radio head (RRH) is Common Public Radio Interface (CPRI). This split architecture has raised an interesting option for future deployments: what if the baseband was centralized and pooled in one location and made to run on standard commercial servers? This concept is called Cloud Radio Access Network, or Cloud RAN. While the idea of centralized baseband is not new, the ability to run baseband function on commercial servers is highly disruptive because it allows operators to separate the hardware platform from the software. This not only reduces the cost of equipment, but more importantly provides additional leverage to operators in deploying and migrating wireless networks that can alter the balance of power with respect to equipment vendors.
Figure 1 Cloud RAN centralizes the baseband processing to improve capacity in wireless networks through coordination among base stations.
We are not yet at the stage to be able to cost effectively run baseband processing on commercial servers. Work is on going to realize this objective in the near future. Additionally, Cloud RAN has to address the question of fiber availability and fronthaul capacity requirements before meaningful deployments can begin (fronthaul is a term that refers to transport of data between the baseband and RRH). These aspects along with detailed regional business case scenarios and financial models were the focus of a recent research report entitled Cloud RAN: Market Analysis of Radio Access Network Evolution that I co-authored with Mobile Experts. But what I like to focus here is the deployment scenarios for Cloud RAN because these scenarios provide new options for wireless network deployment and can be the catalyst for a new cycle of innovation in wireless infrastructure solutions.
The first deployment scenario is to apply Cloud RAN for macro cells. In this case, remote radio heads are the only equipment required at the cell site location. There can be good savings in both capex and opex to be realized in this case because Cloud RAN can reduce site rental expenses, operation and maintenance costs, and energy consumption. Cloud RAN also increases capacity because it enable coordination among cells to reduce interference where multiple base stations transmit and receive from a mobile device: this feature is called Coordinated Multipoint (CoMP) scheduled for LTE Release 11.
Figure 2 Deployment of Cloud RAN for macro cells (Source: 3GPP TR 36.819).
The second deployment scenario is that of a heterogeneous network (HetNet) Cloud RAN where low power remote radios are deployed in the coverage area of a macro cell. This is a very interesting deployment scenario because the level of coordination achieved between the macro cell and the low power remote radios far exceeds what can achieved with a compact base station small cell instead. This leads to higher capacity in Cloud RAN HetNets than small cell HetNets. From this perspective, there is a definite gain to Cloud RAN deployment that brings the cost of capacity lower that it would have been in small cells especially if the cost of deploying low power RRHs or small cells is governed by the logistics and not the equipment. Moreover, this deployment scenario would allow wireless fronthaul to be used instead of fiber because of more manageable CPRI line rate requirements in low-power RRH featuring one or two carrier frequencies.
Figure 3 Deployment of Cloud RAN in HetNet scenario (Source: 3GPP TR 36.819).
The benefits of coordination through CoMP and other interference management techniques like Enhanced Intercell Interference Coordination (eICIC) will depend on a number of factors such as the location on the mobile device, the type of technology (TD-LTE or FDD LTE), and the path (downlink or uplink). The benefits are greatest at the cell edge where coverage is weakest and speeds are lowest. It is not uncommon to realize benefits reaching tens of percent.
The prospect to improve the cost of capacity in wireless networks through Cloud RAN has the potential to define the coming years of wireless networking. Cloud RAN can become a prime engine for developments of new architectures for baseband processing in both hardware (servers and silicon especially co-processors for computationally demanding physical layer tasks) and software (software defined networking/network function virtualization), baseband-RRH interface development, networking and connectivity (fronthaul). This is what makes Cloud RAN an interesting topic to follow in the months and years to come.
About the Author: Frank Rayal's Community Profile
Follow Frank Rayal on Twitter @FrankRayal