This is part of our ‘5G Anyhaul’ blogs series providing expert insights and best practices to help you prepare your transport network for 5G and support innovative services that create new revenue opportunities.
5G promises to deliver higher speeds and support revolutionary new use cases, services and applications that connect people and things. No previous RAN technology has been expected to support such a wide range of services with different capacity, latency, synchronization, reliability and connectivity requirements.
Operators cannot meet these expectations by building the 5G RAN in isolation from other network domains, including the transport layer. Besides the upgrades to the RAN – which include cell densification, more antennas (i.e. massive MIMO) and the use of existing and new operating frequencies – operators need more fiber and new packet transport technologies that address the diverse applications and corresponding network requirements. Where possible, their 5G plans should leverage existing packet networks to save on cost and speed deployment.
5G was not intended to operate independently or replace existing 4G networks. Rather, 5G radios were engineered to complement existing resources. This is evidenced by a thrust of non-standalone (NSA) configurations in early deployments where the 5G radios attach to the 4G packet core network and use 4G LTE for coverage and 5G for capacity infill. Having overlay networks for each RAN generation is undesirable in this configuration. A transport network that converges 4G and 5G transport is more cost effective and simpler to operate.
Embracing Cloud RAN architectures
With the coming of 5G, more operators are interested in moving to flexible Cloud RAN (C-RAN) architectures. C-RAN enables operators to address different application requirements by locating storage and compute resources at the base of the cell site, at centralized hubs hundreds of kilometers away or anywhere in between. For example, operators can support latency-sensitive applications using Multi-access Edge Computing (MEC) data centers that are located closer to the serving cell site.
Figure 1: Flexible placement of RAN functions
This flexibility is enabled through 5G functional splits that divide baseband processing among different elements, including the radio unit (RU), distributed unit (DU) and centralized unit (CU). These splits create two transport segments: fronthaul, between the RU and DU, and midhaul, between the DU and CU.
Development of the eCPRI protocol for 5G fronthaul
In 4G LTE, fronthaul networks rely on semi-proprietary protocols such as CPRI and OBSAI. However, these protocols do not scale cost effectively for 5G because of its use of much larger spectral bands (hundreds of MHz) and massive MIMO. The CPRI cooperation group has created an enhanced CPRI (eCPRI) protocol that scales bandwidth about 10x more effectively than 4G CPRI and thus requires fewer transport resources.
eCPRI is a packetized interface that can be framed within Ethernet to take advantage of the ubiquitous Ethernet networks already in existence. However, due to the time-sensitive nature of fronthaul traffic, which has a one-way latency requirement of about 100 µsec, new technologies are needed to enhance best effort Ethernet to make it deterministic and time bound.
Time-sensitive networking for fronthaul
The IEEE’s Time-Sensitive Networking (TSN) Task Group has published a new standard that addresses TSN for fronthaul (IEEE 802.1CM). This standard enables deterministic connectivity for fronthaul streams within ubiquitous and flexible Ethernet-bridged networks. These TSN Ethernet networks will provide deterministic transport of 4G CPRI and 5G eCPRI streams by controlling traffic scheduling, timing synchronization and system reliability.
Ethernet networks are a shared medium, so it is important to prioritize fronthaul packets over other lower-priority packets. The TSN Task Group has addressed this need with a standard that enables fronthaul packets to preempt other packets (IEEE 802.1Qbu – frame preemption) and keep delays from packet queuing in check.
Figure 2: Avoiding costly overlay networks via converged transport of all mobile traffic
TSN Ethernet networks will provide fronthaul connectivity between the RUs at the cell site and the DUs at the MEC site. As 4G and 5G will coexist, the MECs will serve as 4G/5G C-RAN hubs. But because the 4G radios use the CPRI protocol, the underlying digitized radio, control and management data must be encapsulated onto Ethernet before it can be transmitted over the TSN Ethernet network. New TSN packet switches will perform this encapsulation/decapsulation function in a standardized manner (IEEE 1914.3 RoE) and combine it with 5G eCPRI to maximize transport capacity.
By using these standards-based approaches to enhance Ethernet, operators will get a deterministic network that can address stringent fronthaul requirements. They will also gain the flexibility, traffic efficiency and openness of packet Ethernet networks in a technology that is well understood.
Visit our Optical Anyhaul page to learn how we can help you prepare your transport network for 5G and support innovative services that create new revenue opportunities.
Share your thoughts on this topic by joining the Twitter discussion with @nokianetworks or @nokia using #5G, #Anyhaul, #CloudRAN, #fronthaul, #Optical