Preparing Transport Networks for 5G
The world is going mobile and we’re consuming more data every year, particularly as the popularity of video and music streaming increases, and that’s just from humans. Machines and devices are increasingly acting as consumers of Internet bandwidth and generating even more traffic on the network. Consequently, the industry is in the throes of implementing the 5th generation of mobile Internet technology, or simply 5G, to handle the load, and the anticipated benefits are generating tons of excitement.
However, before we can download Black Panther to our phones in one minute, operators must address the new requirements that 5G imposes, and consequently the challenges for transport networks that will be carrying all those extra packets.
To get there, networks must dramatically increase bandwidth while slashing latency. Accomplishing either one in isolation is difficult, but both at the same time? That requires a new approach, not just doing the same thing better. To do both necessitates splitting the functions in a radio access network. CPRI/eCPRI, IEEE, and xRAN are some of the organizations that have created new standards to design and characterize the new functional splits.
In a recent VIAVI webinar, Dr. Reza Vaez-Ghaemi addressed the challenges of 5G and the implications for transport networks. Reza began by breaking down the service requirements of 5G, and how each poses different requirements for the underlying networks and applications:
- Enhanced massive broadband (eMBB), demands much higher bandwidth availability from the network for the Ue
- ultra-Reliable Low Latency Communication (uRLLC), necessitates extremely low latency in the network design for the relevant network components and their interconnecting transport network
- massive Machine Type Communication (mMTC), requires networks that can serve very large number of end points in a power efficient manner
Reza went on to describe the split options and related service level agreements. Highlights of the session included eCPRI, xRAN/ORAN, IEEE 802.1cm, IEEE 1914.1, network slicing, 5G synchronization metrics, and a discussion of a converged fronthaul network; combining 4G with 5G. Lastly, in order to manage SLAs, service providers need a rigorous testing program.
Reza also recently published a white paper on the same topic.
Q&A from the live webinar:
- For PON in the fronthaul, will a DBA model combining a PON scheduler and RAN scheduler be effective in reducing the upstream delay?
Reza: Yes, it will. RAN and access experts are working on continuous improvements of PON for fronthaul applications.
- Does an Ethernet-based fronthaul network imply more overhead?
Reza: A fronthaul-friendly network implies a network that can prioritize, reserve schedule and deliver the traffic meeting stringent fronthaul SLA’s. This can imply more overhead.
- Is CPRI over Ethernet efficient?
Reza: there are several variants of Radio over Ethernet (CPRI over Ethernet). Some are more efficient, while some others are simpler to implement.
- How do I measure relative TE?
Reza: Measure time difference between two test access points.
- How about more splits on the DU side?
Reza: There are multiple split options; they are illustrated in the webinar diagrams. One slide captures the high-level view of functional split options. Another slide focuses on split options within the PHY.
- Are all the standards already built for a RoE, TSN-based network today? If not, what’s lagging?
Reza: Some standards are approved and public. Some others are in draft status.
- What do you believe are the top three initial use cases for 5G?
Reza: enhanced mobile broadband for example for higher resolution video, augmented/virtual reality, vehicle to vehicle communication (over network), and factory automation.
- What would be the typical requirement from a BC in the fronthaul (i.e. aggregating several RUs)?
Reza: The requirements can be taken from the overall latency and time error SLA’s provided in the webinar.
- How do I manage the synchronization for vehicle-to-vehicle communication?
Reza: The cellular sites communicating with both vehicles need to be sufficiently synchronized.
- What absolute/relative synchronization requirements do you foresee for the “virtualized” implementation of DU instances for LTE/NR?
Reza: The absolute/relative synchronization requirements should be same no matter for virtualized or traditional platforms.
- Are there any challenges with mobility and handover from an area using phase/GPS synchronization in 4G-FDD and 5G NSA 3x to surrounding 4G-FDD cells using normal Frequency synchronization (area with no 5G cells)?
Reza: Yes, there can be challenges, if the user moves between two areas, and they are not frequency synchronized.