By Frank Rayal, Xona Partners and Joe Madden, Mobile Experts
The evolution of base station architecture to a split design comprising baseband processing and remote radio head (RRH) connected by a fiber optical cable heralds the advent of Cloud RAN.
As an architecture, Cloud RAN is much more than centralising baseband processing in one location and using the fibre optical cables to connect to remote radio heads located at the cell sites. More importantly, Cloud RAN is an attempt at decoupling the hardware and software platforms of wireless base stations. Instead of dedicated hardware, operators would deploy commercial servers in data centres to run base station functions and thereby leverage the cost structure of data centres in running wireless networks.
The strategic implications on operator-vendor relationship cannot be more profound as this vision allows operators more agility in implementing network upgrades and selecting between vendors.
Aside from the strategic implications, there are practical and measured benefits to Cloud RAN that fundamentally centre on reducing the cost of network operations. Leaving the cost of equipment aside, there are three main areas for cost reduction: site rental expenses, energy consumption, and network operations and maintenance. These savings are balanced against the cost of the fibre optical network required to meet fronthaul (connection between baseband and RRH) capacity and latency requirements.
After all, supporting 2.5 Gbps for a single 2×2 20 MHz LTE channel will not come cheap. Nevertheless, there are additional benefits in performance improvement to be had. With baseband processing located in a central location, LTE-Advanced techniques such as coordinated multipoint (CoMP) become feasible.
A new report by Mobile Experts and Xona Partners entitled “Cloud RAN: Market Analysis of Radio Access Network Evolution” details the tradeoffs inherent to Cloud RAN with case studies for a North American operator without owned fiber network and an Asian operator with owned fiber network. While the availability of fiber is a prerequisite to a positive business case, determining the savings brought about by Cloud RAN show large variance to regional differences in site acquisition and rental expenses, cost of energy, and network operation practices. Under certain conditions, savings up to 30% in capex and 25% in annual opex can be achieved.
In addition to savings in total cost of ownership, Cloud RAN provides a boost in capacity due to coordinated multipoint. This capacity gain will vary depending on the path (downlink or uplink) as well as on the access mode (FDD LTE or TD-LTE) among other factors (such as the location of the user in a cell as higher gain is achieved at the cell edge). While in some cases the gain is marginal, there are instances of significant gain – on the order of tens of percent – that can be achieved. This provides a further reduction in the cost of capacity in Cloud RAN networks.
As networks evolve to incorporate LTE-Advanced techniques and to leverage the heterogeneous network architecture, Cloud RAN provides an alternative solution to small cells. From a practical perspective, a low power RRH is similar to a small cell, except the baseband being in a central remote location which, in this case, can be a macro cell site. While this has the advantage of greater coordination with the macro cell layer and consequent performance benefit, the disadvantage is in the constant high-capacity requirements for fronthaul (order of Gbps) as opposed to the relatively low-capacity and statistically multiplexed requirements for small cell backhaul (order of 100 Mbps).
Cloud RAN can be viable even if you don’t have fibre assets
Yet, from a pragmatic perspective, the cost of the two architectures can be so similar that Cloud RAN in a HetNet configuration becomes a viable option for wireless network operators even without fiber assets. For this to become reality, wireless fronthaul solutions need to develop to meet the tough requirements for high capacity and low latency.
In the year to come, we expect to see more developments related to wireless fronthaul as vendors will start focusing more energy and resources on this topic which will tend to focus on high-frequency solutions (E-band in particular) as well as on an optimisation algorithm to transport CPRI over wireless.
Areas for development: CPRI, offloading functions
We are still in the early days of Cloud RAN and much work is still required to achieve this vision. For one, commercial servers are not powerful and efficient enough to run certain base station functions. Work is ongoing to solve this problem by innovative specialists as well as established companies (Intel, for example, has been active on this front and their recent acquisition of Mindspeed provides an additional boost).
A popular approach is to offload certain base station functions to a secondary processor similar in concept to the math co-processors installed on motherboards of old PCs. How to do this and what to offload is a space where baseband and silicon vendors will strive to differentiate. Other areas for development include pooling and virtualisation of baseband processing to derive greater benefits and savings from centralisation.
Cloud RAN has the potential to be a disruptive architecture. It has different use cases and application scenarios that almost guarantee it will be a concept to take hold in the future where the business case is amenable. It will also energise new developments to push the limits of a number of technologies as outlined above. With the evolution of LTE to incorporate greater intelligence at Layer 2 and higher layers to expand network capacity, Cloud RAN delivers on performance in a manner the traditional distributed architecture never could.
About the authors:
Frank Rayal is Partner at Xona Partners. He advises investment firms, vendors, and operators on wireless technology and business strategy, spectrum, competitive positioning, and market analytics.
Joe Madden founded Mobile Experts in 2002 and currently works as the company’s Principal Analyst for mobile infrastructure. He has worked in mobile communications for over 20 years and holds a degree in Physics, cum laude, from UCLA.