Raise your hand if you can tell me when 4G LTE was introduced. If you’re curious and look at the search engine results (like I did), you will see that the technology first hit the market around 2010. With that introduction, the networking community began rolling out higher-speed links to cell towers that forced upgrades from copper to fiber to handle the backhaul. With 4G LTE wireless communications, cell sites evolved into small cells as well as indoor and outdoor distributed antenna system (DAS) networks in order to obtain the high speeds enabled by the 4G LTE architecture. Everything worked great. It’s been less than a decade since then. So why change now?
Well, anyone who has been in telecom knows that change is inevitable and a natural evolution of technology. Each time we advance from one generation to another, we reap huge rewards. In the case of 5G wireless communications, the result of this change will be no different. The promise of the wireless leap from 4G to 5G is expected to yield phenomenal improvements. Wireless gigabit speed services with 1-millisecond latency have been being proved out in field trials throughout 2018. The uses envisioned for this new architecture are mind-boggling: 3-D connectivity drones, connected cars, automated traffic control, and so on. So, what about the network? What’s the effect of 5G on the network design?
Effect of 5G
Three distinct sections of the network will see dramatic change as a result of leaping to 5G. These include the radio access network (RAN), the mobile core and the broader network. We will examine both the fronthaul and backhaul elements of the RAN for now. Significant changes are needed to achieve the promised 5G performance. Within the RAN, industry groups are redefining where the computational processing functions of the RAN equipment should be split. This functional split affects everything from physical antenna design to power requirements to fiber placement.
Getting to gigabit speeds and 1-millisecond latency requires higher-frequency wireless spectrum. Higher-frequency spectrum carries higher-bandwidth signals. But physics restricts how far they can travel. That is why 5G requires at least a 10-fold increase (some would argue 100-fold, to max out capacity) in the density of cellular antennas within the same geography. Effectively, this means antennas have to be closer together than existing 4G LTE networks are. For 5G to meet its potential, the distance between small cells must be between 200 and 1,000 feet, according to the white paper, “Paving the Road to 5G with Fiber,” by the Fiber Broadband Association.
In order to adhere to the strict latency and delay requirements of 5G, this essentially requires fiber optic lines be deployed not only to the base station for backhaul to the mobile core but also to the antenna at the top of the tower as a fronthaul connection in the RAN. The standards body that is helping define 5G is the 3rd Generation Partnership Project (3GPP). Where to separate processing functions within the RAN has recently been a topic of 3GPP discussions.
What About Fronthaul?
Fronthaul is associated with the portion of the wireless network connecting the radio equipment or remote radio head (RRH) located at the top of the wireless tower with the radio control equipment or baseband unit (BBU) at the base of the tower. The topic of fronthaul first surfaced for network designers with the introduction of 4G LTE. Backhaul connections transitioned from copper to fiber with the introduction of 4G LTE. At the same time, fronthaul connections began transitioning from heavy coaxial cables to fiber. Fiber optics is now the preferred medium to carry signals in fronthaul. Backhaul is associated with the link from the BBU back into the network. In 4G LTE, the format of the backhaul signal is typically Ethernet riding over fiber optic lines. With fiber on both the backhaul and fronthaul, this requires that significant fiber management and connectivity be planned in base station at the tower.
In many cases, the fronthaul fiber on the tower is protected inside a microduct that terminates inside an outdoor NEMA-4 rated enclosure. This provides a natural transition point for fibers between the antenna and the BBU. Having microduct on the tower for fronthaul not only reduces the cost of installation, but also it lightens the load as well. It can provide nearly 100 times the capacity of broadband over that of a typical coaxial cable, and it has a small footprint. By taking the microduct up the tower (using the traditional hardware), then breaking out the fiber into a distribution box on the tower deck and running the same style of microduct to each radio location, the tower crew can remove all the coaxial cables and replace them with one 10-millimeter microduct (see Figure 1).
The protocol that runs on the fronthaul connection today is called Common Public Radio Interface (CPRI), which resulted from cooperation among Ericsson, Huawei, NEC and Nokia. Together, they call themselves the CPRI Cooperation. CPRI is a circuit-switched protocol requiring a dedicated path, and that means using a lot of reserved bandwidth (up to 2.5 Gbps per connection). The location of the BBU allows a natural test access point in the network for technicians to evaluate the CPRI signals.
5G brings an opportunity to upgrade CPRI protocol to become packet-based, and the result is called enhanced CPRI (eCPRI). This is where many of the technology shifts are affecting fiber placement. Here are some key points about eCPRI versus legacy
· About a 10-fold reduction in required bandwidth
· The eCPRI layer is above the transport networking layer and is not dependent on the actual transport layer topology
· eCPRI is packet-based versus traditional circuit-switched CPRI
CPRI Cooperation has agreed to work for an updated specification eCPRI (2.0). The new specification will enhance the support for the 5G fronthaul by providing functionality to support CPRI (7.0) over Ethernet allowing for CPRI and eCPRI interworking, according to a June 25, 2018, press release from the CPRI Cooperation.
The network topology of 5G will extend the fronthaul concept, and the location of the radio control or BBU is most likely to change. Instead of being collocated at the base of the tower, the BBU function will be brought back further into the core of the network at a centralized location. Centralization of the BBU is a precursor to a virtual cloud RAN. All of this is made possible because of a robust fiber network. Fiber is the enabling technology for the cloud radio access network (C-RAN) (see Figure 2).
Effect of Functional Splits
The trade-offs between computational power and transmission power are at question for network design engineers. What functions exist where? According to analysts at 3GPP and Heavy Reading, seven options are being considered in 3GPP that split the functions of the RRH and BBU computational processing for eCPRI. The actual implementations will vary. It is possible to have heavy processing back in the BBU all connected by the low-latency fiber network with relatively dumb equipment at the RRH. Sounds good, right? But there are tradeoffs. Lengthy fiber connections back to the central BBU mean that more power is required for the fiber transmission lasers. That means that the tower will need more power-generating equipment (see Figure 3).
Designers could move the pivot point around and situate the BBU closer to the RRH, placing more of the computational processing functions in the RRH equipment. That could translate into lower transmission power for lasers. No matter where the functional splits occur, all of this has to work within the latency budget defined within 5G. The one thing you can count on is that more and more fiber will be needed. Fiber is the predominant underlying physical connection for CRAN.
Fiber in the RAN
So, let’s turn our attention to the fiber. The placement of fiber optic lines has moved from interconnecting the core of the network to being deployed closer and closer to the customer. Mass market rollouts of fiber to the home (FTTH), which began in earnest in 2003, signaled the need not only connect to fibers, but also to manage these connections in craft-friendly way such that installers and technicians could connect as many homes as possible.
Early on, splicing fiber was heralded as the best way to achieve low-loss fiber connections in every case. Fiber-splicing is best used when connecting bundles of large-count feeder fiber and long-haul fiber in the network. But splicing fiber at every point in the network proved to be too costly. Splicing fiber requires trained technicians with expensive splicing equipment. In order to reduce costs and speed up fiber deployment to meet the demand for FTTH, the industry has adopted a quicker method of connection that mitigates this expense. The technique is called plug-and-play. The key to making a good connection is in the connector itself, and the target of the industry-leading FiberDeep fiber connectivity products is 0.2-dB insertion loss per connector. This performance is achievable and should be factored into any 5G fiber rollout. We anticipate connecting an abundance of small-count fibers at the edge of the network in 5G and, with FiberDeep technology, plug-and-play is expected to become the leading connection method.
The fronthaul and backhaul fiber network is mission-critical and requires flexible fiber connectivity products designed to offer total configurability for any fiber connectivity issue. Network operators seeking fiber management and fiber protection turn to Clearview Cassettes to acquire optical components and small-port-count fiber terminations as inexpensively and conveniently as possible. These products allow technicians to maintain bend-radius protection while working in a small space. Bend-radius protection is just one measure that supports full throughput of the fiber signal. Additionally, the cassette technology lets service providers isolate one fiber sheath from another in the same enclosure, 12 fibers at a time.
These fiber management and fiber protection practices can help reduce not only the deployment time but also the operations expense of running the fiber network once installed. In general, plug-and-play technology coupled with the scalable fiber management of the Clearview Cassette and fiber protection enabled by microduct simplifies fiber installations and maintenance, allowing providers to reduce costs, maximize resources and turn up subscribers more quickly.
The 5G cell site density and topology are shifting dramatically in favor of all-fiber connections. The implications are that higher-bandwidth signals ride best over fiber, and 5G is going to be flooding the airwaves with these signals in just a few years. Getting the wireless signals onto fiber as soon as possible is required in order to achieve the promised 5G performance. 5G network designs dictate the use of more fiber, which means increased needs for plug-and-play, fiber management and fiber protection throughout the RAN.
“Industry Leaders Releasing New Functionality for the eCPRI Specification for 5G – eCPRI 2.0”;June 25, 2018; press release from the CPRI Cooperation. www.cpri.info/press.html.
“Paving the Road to 5G with Fiber,” by the Fiber Broadband Association. www.fiberbroadband.org/page/paving-the-road-to-5g-with-fiber.
Kevin Morgan is chair-elect of the Fiber Broadband Association and chief marketing officer of Clearfield, a supplier specialist in fiber management and connectivity platforms for communication service providers.