By Tammy Parker
One of the hottest industry acronyms these days is C-RAN, which is dominating the advanced wireless markets of Asia, where fiber is abundant and high population densities make the architecture fiscally beneficial. But proponents of the technology believe it could eventually spread across the globe.
C-RAN stands for Cloud RAN (radio access network) as well as the less catchy Centralized RAN. In either case, C-RAN leverages distributed base station architecture to enable a host of benefits, such as capex and opex savings, increased asset utilization, and savings on energy.
According to Nick Marshall, research director at ABI Research, the Centralized RAN concept separates the radio and antenna parts from the digital baseband path and pools multiple baseband units (BBUs) in a central office, or base station hotel. These digital-only base stations are linked via fiber to remote radio heads (RRHs).
Building upon this concept to create a Cloud RAN, an operator can use Network Functions Virtualization (NFV) to shift digital functionality resident in the base station hotel to the cloud. Baseband tasks would be accomplished through the use of general purpose processors and generic server farms, which would link via fiber to a virtual server somewhere in the cloud, Marshall added.
Intel offered a look at a cloud RAN architecture in a recent white paper.
Similarly, Guang Yang, Strategy Analytics senior analyst for wireless operator strategies, said: "If we move further from the centralized BBUs, they can be virtualized based on general purpose IT architecture," creating "the real Cloud RAN."
"This type of cloud and virtualized RAN would change the RAN ecosystem significantly. IT companies like Intel and IBM could become the hardware suppliers for RAN infrastructure," Yang said. "It would be a big challenge to current leaders in RAN infrastructure market like Ericsson and Huawei. So they oppose it strongly," he added.
Yang observed, however: "I don't think the real Cloud RAN will be massively deployed in near future."
Benefitting from a new approach
Sensing possibilities to broaden its influence and market share, Intel has been involved in C-RAN since it was first conceived and is the only major chip maker supporting the technology. The company began working with KT during the fall of 2010 as the South Korean operator drew up plans for its Cloud Communications Center (CCC) architecture.
This was a precursor to C-RAN work that Intel began with China Mobile the following spring. China Mobile has conducted numerous C-RAN trials and is expected to deploy C-RAN on its commercial networks between 2015 and 2016, according to ABI. Intel has also worked with SK Telecom--which announced a virtualized RAN structure called SCAN, for Smart Cloud Access Network--during October 2013.
Intel and other C-RAN supporters cite numerous advantages from the architecture. According to Sandra Rivera, director of Intel's market development, communications and storage infrastructure group, China Mobile has estimated a 30-60 percent savings in total cost of ownership from C-RAN architecture.
The figure is closer to 30 percent if an operator does not have much fiber and needs to deploy it to support the C-RAN, Rivera noted. But an operator with a substantial fiber footprint can generate 60 percent savings by deploying C-RAN.
China Mobile detailed a trial in Zhuhai that shows that, compared with traditional RAN deployment method, C-RAN centralized deployment can reduce "the TD-SCDMA network's capex and opex significantly."
Rivera said that many wireless network assets sit idle for as much as 70 percent of the time. "They have to be installed and dimensioned for peak traffic, but when people leave urban areas and head home in the evening, those assets sit idle," she said. But with C-RAN, "you don't have to over-dimension your infrastructure when you deploy," Rivera continued.
Further, C-RAN enables easier network extension, as an operator can install new RRHs and link them to the base station hotel as needed.
C-RAN slashes capex because fewer BBUs are needed, which reduces opex because fewer BBUs means less energy consumption and diminished maintenance costs. The reduced energy consumption makes C-RAN a "green" alternative, with China Mobile estimating 71 percent power savings vs. traditional RANs.
In addition, Yang observed that centralized BBUs provide latency improvements and can improve the performance of LTE Advanced coordinated multipoint (CoMP) or other collaboration mechanisms significantly.
C-RAN also can deliver new revenue streams by enabling an operator to introduce new levels of programmability and capabilities in place of what was "a static, fixed-function approach at the base of every tower," Rivera said.
Getting in front of fronthaul
Fiber, needed for fronthaul, is crucial to C-RAN deployment, so it is no wonder that fronthaul is constantly brought up as Cloud RAN's biggest challenge. Fronthaul connects RRHs to the aggregated BBUs, with traffic then backhauled from the BBUs to the IP core or evolved packet core (EPC).
"The deployment of C-RAN requires many dark fibers for fronthaul, or operators have to deploy many new fibers," Yang said, adding that "only those operators with plenty of fiber resources can deploy C-RAN in large scale."
This, he noted, has so far restricted C-RAN deployments to fiber-rich Japanese, Korean and Chinese operators.
Operators use optical network (OTN) protocols to manage the large fronthaul fiber network needed for C-RAN. In April 2013, South Korean operator LGU+ hopped on the C-RAN bandwagon when it announced it would use Huawei's wavelength division multiplexing (WDM)/OTN technology for common public radio interface (CPRI) fronthaul on the LGU+ LTE-A network.
"The CPRI-over-WDM/OTN solution enables pooled baseband processing for distributed base stations, saving fiber resources and improving network reliability. The solution is a typical application of the 'WDM/OTN-to-the-edge' idea in LTE backhaul scenarios," the firms said.
Yang noted that operators without large fiber footprints can resort to "local centralization," by centralizing a few neighbor base stations in a limited area where better performance is needed due to high traffic volume.
"We can see Ericsson's Radio Dot as this kind of C-RAN. The baseband of the Radio Dot system can be shared with the nearby macro base station. It actually forms a centralized baseband pool to support tight coordination between outdoor and indoor coverage," Yang added.
Other ways to get around the fiber conundrum include CPRI compression, wireless fronthaul and bi-directional fiber, Intel's Rivera said.
Regarding wireless fronthaul, Ceragon has introduced a "virtual fiber" wireless product, the FibeAir IP-20C, for backhaul and fronthaul. At the 2014 Mobile World Congress, the company conducted a demo employing CPRI Option 3, with an eNodeB connected to RRHs using wireless fronthaul.
There are also efforts to move beyond CPRI for C-RAN and other HetNet deployments. In C-RAN, transport between the baseband module and remote radio module usually relies upon CPRI. However, because CPRI does not guarantee interoperability, ETSI is working on an Open Radio Interface (ORI).
ABI's Marshall noted that CPRI standardizes the data plane but lets the control plane remain vendor-specific. ORI will build on CPRI and standardize the control plane, enabling operators to mix and match equipment from various vendors.
"That's going to be important as operators roll out hundreds of thousands of remote radio heads and HetNets, so they can mix and match for capacity enhancements or dual-sourcing or whatever reason," he said.
According to Yang, "CPRI was actually driven by vendors, but ORI is mainly driven by operators. So, ORI makes more sense for operators."
C-RAN vs. small cells
One question that frequently comes up is whether C-RAN competes against small cells.
Yang feels they are complementary and suited to different deployment scenarios. Smalls cells will be more cost effective in homes and small offices, while C-RAN may be more efficient in terms of capacity and performance for medium- to large-size buildings and public venues like airports and shopping malls.
In addition, a whitepaper from SOLiD noted that multiple RRHs and small cells can be fronthauled on a fiber ring to provide redundancy.
Marshall noted that the terms "C-RAN" and "small cell" are sometimes used interchangeably. For example, SoftBank's Wireless City Planning subsidiary in Tokyo operates what it claims is one of the densest small cell networks in the world, but "it is all actually C-RAN," Marshall said, delivering TD-LTE over former personal handy-phone system (PHS) frequencies.
Another uncertainty regarding C-RAN is when, or whether, C-RAN will break out of Asia.
"What remains an open question in my mind is the Western economies and countries. These places have fiber, so what's the hold up?" Marshall observed.
Regardless, C-RAN is already helping drive RRH shipment volumes. ABI forecasts that in 2014, half of RRH shipments will be LTE-based. By 2018, the total number of shipments will amount to more than 2 million units, almost double the number shipped in 2013, the firm said.
MarketsandMarkets predicts the global C-RAN market will grow from $1.71 billion in 2013 to $11.31 billion in 2018, representing a compound annual growth rate (CAGR) of 45.9 percent.
Article updated May 1, 2014, to reflect the fact that Guang Yang is with Strategy Analytics.