It will take operators ten years to achieve full virtualization of radio access networks once they begin the process, and delaying this process further will have the knock on effect of putting 5G back by a few years, ReTHINK Technology Research says in its Cloud RAN Deployment Forecast 2017-2025.
Fully virtualized Cloud RAN is being held back mostly by the lack of interoperable standards. The standards that exist, for instance in front haul, are controlled by a handful of players, who have no vested interest to make them fully interoperable, instead they place proprietary flavors on them keeping prices high.
Centralized or Virtualized Site Installed Base Versus Traditional
This is one of the main things preventing C-RAN from happening earlier, and most shipments will not start until this has been ironed out, a process that could take another 3 years before shipments break the 3 million a year barrier.
As a result, centralized and virtualized macrocells and microcells will be deployed at a compound annual growth rate (CAGR) of 23 percent between 2017 and 2025 and although these will overtake new deployments of conventional cells in 2019, most of these deployments will still be centralized rather than fully virtualized. Once operators start to deploy RAN VNFs, usually conforming to ETSI NFV specifications, there will be a sharp acceleration of growth in the number of Cloud-RAN sites, from 2020.
There is now a major dilemma for operators between simply using Centralized-RAN or going for a fully Virtualized RAN. The former is simpler and delivers some efficiencies; the latter is very difficult and disruptive to implement, but we believe will revolutionize cellular economics. Full Cloud-RANs will remain rare until mid-2020.
The two key barriers have been availability of fiber and the vendor stranglehold on the CPRI (Common Public Radio Interface) interface. That has meant most early C-RANs have had to be very localized. It is essential that new emerging standards, preferably from the IEEE, take hold to open up the ecosystem and accelerate adoption.
Operators which support full Cloud-RAN will have a far easier, cheaper task to implement 5G. However, the process of virtualizing the RAN may delay 5G upgrades.
By 2025 we shall see the start of the fully hosted RAN – virtualized on a third party cloud (a vendor’s or perhaps Amazon’s) and this will transform the economics and power base of the entire mobile industry.
The biggest barriers to all types of C-RAN are related to fronthaul, mainly the availability of affordable, high quality fiber and the need to use the CPRI interface, which has been controlled by a small group of vendors, which implement it in semi-proprietary ways. This raises high barriers to entry for alternative equipment providers, such as those supported under the Facebook Telecom Infra Project and threatens MNOs with lock-ins and high prices. It also makes many C-RANs economically non-viable because of the high cost of low latency fiber and CPRI.
So far C-RANs have been relegated to the indoors, such as arenas and airports, using shorter fronthaul or mid-haul links, which use Ethernet. The right way forward is the release of cheap dark fiber, and so far this has only happened in Japan. Some other countries, such as the UK, are even trying to get by on copper-based technologies such as G.fast. We expect copper to grow from 4 percent to about 16 percent of fronthaul connections between now and 2022, and then to decline as fiber emerges.
The bulk of long range fronthaul will be on fiber and the main change will be the rising use of dark, rather than leased, fiber as operators acquire the skills and tools to light up the fiber themselves.
The promise of cloud radio access network (C-RAN) technology has always been great, but what was needed was the support of one of the major radio and base station original equipment manufacturers. After all, for the last three decades or so, mobile operators have purchased outdoor RAN equipment for a specific market from a single vendor, because the interfaces between the various RAN components are not generally open.
When small cell architectures were first discussed, many believed that those architectures would open up the RAN to other vendors. But when it came to implementation, the same proprietary interfaces were still there, and the outdoor small cell vendors were all forced to look for new markets in-building. As a result, the Long Term Evolution (LTE) RAN market consisted of the same RAN vendors. After a series of mergers and acquisitions, we are now left with five vendors who operate globally: Ericsson, Huawei, Nokia, Samsung and ZTE. All are stronger or weaker in various markets, and the vast majority of operators buy RAN equipment from one or more of these vendors. Some operators (usually operating in rural markets) have deployed other vendors for specific markets; Vanu is a good example.
Virtualization of the mobile network has started with the core and slowly expanded outward toward the RAN. The RAN is generally seen as the last part of the network to be virtualized, with the deployment of a virtual baseband unit (vBBU) and radio using off-the-shelf components with open interfaces. Just as virtualization has reduced the cost of deploying and maintaining an Evolved Packet Core (EPC), so the goal is the same for the RAN — reduced capital expense (capex) and operating expense (opex).
xRAN, a membership organization named after extensible RAN (xRAN), provides a good example of the current work going on in RAN virtualization. The goal of the industry group is simply to develop, standardize and promote a software-based, extensible RAN and to standardize critical elements of the extensible RAN architecture. xRAN members include some of the biggest and most advanced operators in the industry: AT&T Mobility, Deutsche Telekom, Telstra, Verizon Wireless and SK Telecom. Note that these operators have been pushing virtualization hard and are also moving ahead to 5G as quickly as possible. xRAN vendor members include Intel, Cisco, Mavenir, Amdocs and others.
What has been missing industry groups involved in the open C-RAN debate is the involvement of the big RAN OEMs. Without one or more of the big vendors willing to move to an open C-RAN architecture, there is little chance of getting C-RAN deployed meaningfully into the major markets for outdoor cells.
So now the big news: Nokia has joined xRAN. The Finnish company has been working behind the scenes for a few months and has now executed all the necessary paperwork (and, I assume, has written a check for the dues).
When I discussed this news with someone in the industry, their first reaction was skepticism, because the major OEMs have joined similar virtualization and open forums in the past, only to use the opportunity as a fact-finding exercise without making any changes in their strategy or the openness of their products. But having spent a week in December in Finland at Nokia’s industry analyst event, I concluded that Nokia is sincere: The company has made a big move to cloud architectures using open interfaces, and xRAN is the latest development with this strategy. In short, it does not appear that Nokia is simply in this to sit back and listen, but to contribute to the forum and move toward open RAN interfaces as quickly as possible.
It is also worth remembering that building, optimizing and operating radio networks is always harder than it appears. Although some people draw comparisons to Wi-Fi, LTE and soon-to-come 5G are completely different animals. 5G will support network slicing and prioritization of traffic. All cellular networks hand off between cells to (hopefully) maintain the connection. These all add complexity to the network and operators. As such, Nokia and its major OEM competitors have considerable experience and expertise building and operating networks, expertise that will be as valuable as it has ever been as the industry moves to 5G.
Nokia is unlikely to lose its place in the industry simply because the company is moving to open RAN interfaces. With open architectures and virtualized 5G networks, there are considerable opportunities for network analytics, optimization and professional services. There will be no shortage of things to do.
The question now is, assuming this initiative is successful and continues to make progress, what will Samsung, Huawei, ZTE and Ericsson do. Nokia’s move has put pressure on the other OEMs to follow suit, if not by joining xRAN, then at least by demonstrating a viable, open RAN architecture. 2018 is going to be interesting.
Read the rest of the March AGL Magazine HERE.
Iain Gillott is the founder and president of iGR, a market strategy consultancy focused on the wireless and mobile communications industry. The company researches and analyzes the effect new wireless and mobile technologies will have on the industry, on vendors’ competitive positioning and on its clients’ strategic business plans. Visit www.igr-inc.com.
Wells Fargo Securities Senior Analyst Jennifer Fritzsche published this Equity Research note recently on fronthaul and centralized radio access networks (C-RAN) and their impact on the wireless carriers, tower companies and fiber providers.
“Simply put – we see this trend as one in the VERY early innings. In terms of implications for specific sectors on our list: 1) Wireless Carriers – positive for network capacity and depth; but it will likely require capital to continue to be spend; 2) Towers – Neutral/Slightly Positive – Less space is required at base of tower, but more equipment and fiber would be placed on the tower. The shift in capital spend could be directed more to C-RAN and less to macros; 3) Fiber Companies – A significant positive, in our view. Fiber is the critical element in this initiative. Carriers need it to carry out C-RAN’s promises and capabilities.
“WHAT IS FRONTHAUL? Fronthaul plays a critical role in C-RAN, an architecture that places a baseband unit at a centralized location, with the ability to serve multiple remote radio units (RRUs) attached to wireless towers. The centralized baseband units and RRUs are typically connected by a fiber-optic line referred to as “fronthaul.”
“WHAT DOES THIS MEAN FOR WIRELESS CARRIERS? C-RAN architecture can provide a host of benefits to wireless carriers seeking to densify their networks, such as savings on capital and operating expenditures and lower power consumption needs. From a capex perspective, fewer baseband units are needed to service multiple remote radio heads. However, the fiber company must be paid. As a result, it is hard to assess the total capital outlay and how it may compare to “traditional” wireless capex spend.
“TOWERS – NEUTRAL EVENT FOR MOST/POSITIVE FOR Crown Castle International (CCI) – The impact to the tower companies from C-RAN deployments is less clear. On one hand, wireless operators will not need as much ground space at the cell site to house baseband units. On the other hand, the tower companies should see more leases for remote radio head attachment as these networks are built out. The best positioned tower company to capitalize on this trend is CCI given its Sunesys acquisition, in our view. This acquisition gave CCI ownership or rights to an additional 10,000 miles of dense metro fiber. This transaction not only allows CCI to control some of the connections to its 14,000 small cell nodes, it also enables CCI to provision dark fiber fronthaul as C-RAN topologies expand. Many believe CCI could be more active in the M&A space, with Sunesys as the first move for the company in the fiber space.
“FIBER COMPANIES MAY HOLD KEYS TO THE KINGDOM – Dense metro fiber networks for fronthaul are a critical component to make C-RAN deployments economical. The fronthaul can take many forms, such as carrier Ethernet, wavelengths or microwave, although dark fiber is the most efficient for large-scale deployments. ZAYO is best positioned to play here, in our opinion, given its large exposure to the wireless space already. Level Three Three (LVLT) – while it typically shied away from fiber-to-the-tower (FTT) builds – certainly has the size, scale and assets to be a serious player here if it wanted.”
— Jennifer Fritzsche
October 23, 2014 — The incorporation of small cells, DAS, remote radio heads and Wi-Fi (known collectively as HetNet) into mobile networks is bringing with it a heavy layer of complexity to the backhaul network and higher costs.
To deal with that, software-defined networks (SDN) and network functions virtualization (NFV) solutions are being use to provide greater backhaul flexibility and cost-savings. The key here is cost-savings. NFV addresses moving from an environment that is hardware-centric and porting it to software network architecture. NVF also facilitates fast times to market for products. NFV is a carrier-driven platform and they have a huge interest in it.
An iGR Research study released during the summer examines the cost savings, both capex and opex, that mobile operators can expect when moving to SDN, NFV or cloud radio access network (C-RAN) architecture.
“The RAN and evolved packet core (EPC) are slowly becoming applications that can run on off-the-shelf IT infrastructure hosted by data center operators and other third parties,” Ian Gillott wrote.
First, current architectures where the baseband units (BBU) are deployed at the base of the cell tower must change to an architecture in which the BBUs are deployed in one (or more) centralized data centers, according to iGR. Second, the baseband functionality itself should run in virtualized software on generic computing platforms.
“This eventual architecture would divorce the 1:1 ratio between baseband and RRH, and thus give mobile operators the ability to support the same number of cell sites and sectors on less hardware. This would therefore be a true C-RAN deployment in a commercial data center,” Gillott wrote.
iGR’s market study, “Cost Considerations for Centralized RAN and Cloud RAN,” can be purchased and downloaded directly form iGR’s website: igr-inc.com
Ernest Worthman is the editor of AGL Small Cell Magazine.
J. Sharpe Smith, AGL Small Cell Link editor, contributed.
July 24, 2014 — For the last few years, the radio access node (RAN) camp and the small cell camp have stood on their respective podiums, maintaining they are competing technologies with different architectures. Well, it seems the warring camps have made peace.
Recently, it appears that the cloud — the latest and greatest virtual network platform for everything and anything — is being married to a set of seemingly disparate platforms. As is now being proselytized, this union may just be the best option for the next generation of high-density networks by utilizing baseband pooling and distributed baseband, and enabling LTE Advanced features within these networks.
But wait a minute…small cells offer high-density traffic solutions by using baseband distribution – the exact opposite of C-RAN. So, confused? Me to! So let’s see if we can make sense of this.
C-RAN is a marvelous idea. The idea is to reduce operational cost, by consolidating everything in one location. That has a number of advantages:
• It allows the technician to work more efficiently at maintaining baseband resources.
• It enables easier implementation of LTE-Advanced features, which improve capacity.
• It improves performance by placing multiple RRH units on a single baseband node.
• Financially it is attractive through the use of technologies such as Coordinated Multi-Point (CoMP), which improve throughput by as much as 80 percent at the cell edge for both uplink and downlink.
• Additional savings come from the centralized location of the baseband processing, making upkeep simpler and more efficient.
But wait…small cells are a marvelous idea too! Small cell architecture is designed to reduce the cost bit by distributing the baseband, reducing the cost of the radio hardware and increasing spectral efficiency and overall throughput.
There are other metrics involved in this, of course, but the advantage of C-RAN is that it implements virtualization in a base station system, allows for dynamic allocation of processing resources within a centralized baseband pool to different virtualized base stations, and different air interface standards. This allows the operator to efficiently support the variety of air interfaces, and adjust to the ebbs and tides of different deployment and bandwidth demands. At the same time, the common hardware platform will provide cost effectiveness to manage, maintain, expand and upgrade the base station.
Truth be told, it reads about the same for small cells. Virtualization will be implemented there too, eventually. Well, now we know, and the marriage looks pretty good to me.
Ernest Worthman is the editor of Small Cells magazine.