5G is bringing higher bandwidth to mobile users, which pushes more data on the optical networks. Optical networks need greater capacity and more flexibility.
Every mobile phone user expects 5G to let them download movies faster and more reliably than LTE networks. Plus, 5G will open the door to new use cases we haven’t conceived yet. The greater demand for data puts pressure on optical networks to not become the bandwidth bottleneck. Furthermore, optical networks need the ability to adapt to changing demand. 5G Technology World spoke with Nathan Tracy, TE Connectivity and Optical Internetworking Forum (OIF) President; Jonathan Sadler, Infinera and OIF Networking & Operations Interoperability Working Group Chair (acting) to learn how optical networks are adapting at both the physical layer and the control layer.
Tracy: 5G will drive significant increases in network capacity. The market projections indicate that 5G will increase demand for optical communications around the world, though it will differ by location. In the U.S. carriers such as AT&T and Verizon are focusing on mmWave, which will need lots of fiber to deploy.
Sadler: I come from the perspective of network control–the setup and breakdown of calls through the network. The radio access and fronthaul part of the network is relatively stable. It’s in the backhaul where things get dynamic. That changes depending on what people want from the network. For example, voice calls or internet access.
Traffic patterns can quickly change in response to planned and unplanned events. This isn’t the old voice network where you knew calls would increase on Mother’s Day. You could plan for that peak in demand. Now, things change hourly, especially where when dealing with streaming video. With 5G, the connection to the user is so much faster and the potential for bandwidth crashing couldn’t be greater. How do we deal with the dynamics of changing bandwidth demands between endpoints? We’re no longer dealing with 64 kbps data rates, but multi-megabit data rates.
Sadler: Once you get beyond the fronthaul, that is, the first hop into the network, there’s a lot of variability as to how people will build their networks. While the radio access network is static, the rest depends on where the network starts making traffic decisions.
Tracy: The network architecture is being defined to give operators and equipment developers flexibility as to where the various network functions are being located. For example, if an operator’s cost for locating a cell is expensive because of the cell’s size, height, or weight, the operator might choose to install the equipment at another location. In the case of a latency-sensitive applications, the mid-haul and backhaul could be in different locations. The locations could depend on economical, physical, or operational issues. 5G network standards provide the flexibility to decide on locations and integration levels. It’s difficult to say, even within the fronthaul because even radio access networks (RANs) could be virtualized. It’s hard to imagine the use cases that might arise because of the flexibility built into the standards.
Tracy: Recall that the Ethernet specification defines a physical layer and a MAC layer above it that is used to transport packets. In terms of the physical layer, it’s going end-to-end.
Sadler: As for the Media Access Control (MAC) layer, it depends on the applications in the layers above. Network operators can choose where to put that boundary. The Ethernet physical layer has volume and that lowers cost.
Tracy: We expect the optical network to change from using wavelength-division multiplexing (WDM) to coherent optics. The OIF recently passed a 400G(bps) standard calling for coherent optics that will come from different suppliers and be interoperable. We will no longer need “bookended” coherent deployments. Now, the OIF is starting to ask “What’s next beyond 400G? That’s led to an effort to survey network operators on what they expect to need for optical data rates. We recently attended industry events where we held meetings with network operators. Our meetings helped to tailor the survey. We’re now ready to get into the details on what should be the next OIF project relating to coherent optics. There are tradeoffs. For example, we can go for higher data rates or for longer reach or the ability to make that trade off on the fly.
Flexible Ethernet (FlexE) has gotten some traction in terms of 5G and network slicing. Some networks will use FlexE to allow for network virtualization.
Tracy: The answers vary depending on time frame. We’ve heard interest in 800G, but we need for the network operator group of OIF to do their job and come to the right conclusion. Looking beyond, we’re hearing numbers such as 1 Tbps and 1.6 Tbps. With regard to reach it’s possible that the next solution could address both bandwidth and reach where you get longer reach if you’re willing to roll back to a lower bandwidth. Hypothetically, there could be applications for 800G but where the speed drops to 400G for longer reaches. It’s still too soon to know how that will settle out.
Sadler: That discussion is also impacting the dynamic, control systems side of the network. Operators need to understand the impacts of new optical links and how they affect network operations. That leads to a discussion of where to place the interface between optics and electronics and how that affects an increase or decrease in demand. Links could change speeds in response to demand.
Sadler: A disaster event could cause a sudden reduction in network capacity, forcing a workaround. A disaster in another part of the world such as an earthquake that people want to learn more about can cause in increase in demand. It could be a planned event such as a sporting event. Sports stadiums are empty more than they’re filled. Do you have enough bandwidth available during the idle periods or do you bring up capacity just for games? 5G will result in situations where when the stadium is filled, there will be many high-bandwidth users. 5G will bring new abilities and we don’t yet know how those abilities will be used.
Sadler: Definitely. People who may not have high bandwidth internet in their homes are now being forced to use wireless. I heard a story about a family that had a single iPad connected over LTE. Where previously the iPad had casual use, now three kids needed it for school. The current coronavirus situation has caught people off guard in terms of equipment, but it’s affected network operators too. Forecasts are good because they provide ideas as to where to put network equipment, but forecasts could never have predicted the current situation.
Netflix must be happy that they switched to a virtualized streaming delivery service, which provides better flexibility. A few years ago, Netflix switched from using its own equipment to using a cloud service provider. That let’s Netflix turn up and bring down services as needed. They can bring up use of their back-end video servers to meet demand. That demand also drives the need for using the optical network.
Tracy: We’ve all heard about the “edge cloud” in relation to 5G, which will drive the need for edge computing. While the actual computing may be done by cloud service providers, the mobile network operators may use that capacity as part of network virtualization.
- OFC plenary: 5G will drive optical
- Industry’s First Terabit-Scale Ethernet PHY Enables Highest-Density 400 GbE and FlexE Connectivity
- An Overview of 3GPP Defined OTA Testing Methodologies for 5G Devices
- Cloud-native 5G data suite supports automated network processes and data monetization