When it comes to 5G, it’s hard to skirt an undeniable truth – this revolutionary technology has a chicken and egg problem.
To be economical for MSOs and carriers to roll out, 5G needs widespread adoption of revolutionary technologies like autonomous vehicles, virtual reality, and the Internet of Things (IoT). However, these technologies and applications themselves need robust 5G networks for their own development and testing.
The pace of change is only picking up, however. With big-name mobile operators developing, launching and expanding their own 5G networks, the future of this revolutionary cellular standard and the innovative technologies it will support is bright.
From O-RAN and evolving mobile fronthaul architectures to network slicing, MSOs, carriers and other service providers involved in 5G are now facing new hardware, interoperability and network design needs. In this blog, we examine some of the top issues network operators are facing around 5G deployments and how we can help.
From 4G to 5G: Evolving Network Architectures
Let’s go back in time for a moment and examine a conventional distributed (D-RAN) 4G RAN architecture. In this scenario, interconnected base stations, eNode-Bs, each feature a Remote Radio Unit (RRU) at the top of the tower and a Base Band Unit (BBU) in a cabinet at the bottom. The BBU operates as the interface for backhauling traffic to the core network. The RRU and BBU both leverage proprietary hardware.
Another important aspect of this conventional 4G D-RAN architecture is the Common Public Radio Interface (CPRI), the protocol for BBU-RRU transport, connectivity and control specifications that was first established in 2003. CPRI relied on having the BBU house the physical layer (PHY), data link layer and network layer architecture. Fronthaul connectivity connected the RRU with the BBU’s PHY layer. CPRI’s delay budget meant that the RRU and BBU could not be located too far apart from one another.
However, as the industry evolved towards widespread adoption of 5G, network operators needed a solution for the inefficiencies of CPRI. While a decent protocol for 4G, the industry quickly realized that CPRI simply could not scale upwards to support 5G’s performance requirements. CPRI’s need for a dedicated link for each antenna is incompatible with 5G’s massive multiple input, multiple output (MIMO) technology. Solving that required evolving the functional split between the RRU and BBU and disaggregating this traditional pairing into multiple components. And what was the end-result? Flexible architectures that network operators can use to optimize 5G fronthaul to support the plethora of different end-user use cases and applications coming down the pipeline.
Moving from CPRI to the new enhanced CPRI (eCPRI) technology also resulted in a shift in fronthaul requirements from 10G with traditional 4G LTE to 25G with 5G. While eCPRI is 10 times more efficient than CPRI, 5G is 100 times faster than 4G and will handle much more throughput, supporting up to 1 million devices in a square kilometer. As a result, 5G has an augmented bandwidth requirement. Whereas links up to 10G used to be enough capacity for 4G (in a previous blog we noted that CPRI speeds ranged from 600Mb/s to 10Gb/s), 5G fronthaul requires link capacity up to 25G. As a result, demand for 25G optical transceivers will only get stronger as more carriers and MSOs roll out 5G networks around the world.
A Word on Open RAN
Enter 3GPP TR 38.801, the IEEE and ITU-T. As we’ve written about before, this model has opened up the RAN into three main components: a Radio Unit (RU), a Distributed Unit (DU) and a Centralized Unit (CU). The fronthaul connection exists between the RU and DU, while the link between the DU and CU is characterized as midhaul. The network connection between the CU and the core becomes the new backhaul environment.
This new architecture built upon earlier, gradual moves (Cloud RAN, Virtual RAN) to open up the RAN to Commercial off the Shelf (COTS) equipment and vendor-neutral software-defined technologies. And now, the industry has taken another important step further. Whereas CRAN and VRAN mostly concentrated on moving and virtualizing the BBU, modern Open RAN has taken the idea of disaggregation to the next level with the RU, DU and CU components. Being open source in nature, Open RAN designs empower network operators to mix and match equipment and software from different vendors in order to create and deploy the 5G architectures that will best suit their needs.
Here, the deployment options are many. When it comes to 5G scalability and cost-effectiveness, location is everything, and 5G faces a special conundrum. While its wavelength range is shorter than 4G, 5G will support many more devices and, by extension, significantly more data. As a result, network operators are looking at creative ways to deploy 5G infrastructure. While standard macro cells on top of buildings and cell towers will still be in vogue, operators are also leveraging other unique installations such as pole and strand mounts. Being able to leverage cell site, split, dual split, remote CU-UP, and central CU-UP RAN designs adds another layer of customization, flexibility and complexity into modern 5G rollouts.
Network Slicing: A Path Forward for 5G
Network architectures are only one side of the equation, however. Think about all the potential use cases (we know about) for 5G: mobile broadband, ultra HD content streaming, gaming, machine-to-machine communication, autonomous vehicles, robotics, remote surgery, etc. Per SDX Central, each of these use cases requires a “unique set of optimized resources and network topology – covering certain SLA-specified factors such as connectivity, speed and capacity.” Flexible deployment architectures can help speed the proliferation of 5G, but the real proof of 5G will be in how it meets various application requirements.
Network slicing is arguably a key piece of the 5G puzzle. The basic premise of network slicing is the creation of multiple virtual networks that share the same physical infrastructure. This technique effectively supercharges network agility and the affordability of deploying 5G. Instead of offering every market segment full functionality at once, network capacity and other resources can be dynamically assigned as needed. For example, applications that leverage mobile broadband can be assigned their own slice or virtual end-to-end network, while other applications around IoT can leverage another. Each slice is customized to meet the exacting requirements of that application. Network slicing thus speeds 5G rollouts by lowering the capital expenditures needed upfront.
The Precision OT Value
New network architectures and virtualization techniques like network slicing are helping network operators speed the deployment of 5G. At the same time, they’re also posing a myriad of new challenges for service providers around interoperability, aggregation and data transport, and installation of cell sites.
At Precision OT, we see ourselves as the glue that can help network operators piece all the different components of their 5G networks together for seamless network operations. With our customers, we wear many hats. On one hand, we can help them design a network and pick the right components for 5G fronthaul, such as our line of 25G DWDM Fixed, Tunable and Bidirectional transceivers with ranges between 10km and 40km. These products have different offerings including Industrial Temperature ranges to overcome the outdoor environmental challenges of RAN deployments. On top of the different transceiver options, we also offer a line of passive optical components such as MUX/DEMUX solutions that can fit your 5G specific deployment needs.
More importantly, we offer expertise in systems interoperability. Backed by a testing lab and a team of innovative engineers (check out our Advanced Engineering Group here), we can help our customers ensure that all the pieces of their network will work as desired. We’re on a mission to end vendor lock, but that’s best accomplished with expert help and trusted partnerships. We’re that partner.