Skip navigation
Like what you’re reading?

What are the latest findings and trends in packet fronthaul?

Packet-based fronthaul solutions have been picking the interest of CSPs worldwide in the last few years. However, while the journey of packet fronthaul has started, its use for centralized RAN and Cloud RAN deployments is still evolving. So, what are the latest trends and findings in the area? Here is the first-hand experience.

Strategic Product Manager


Strategic Product Manager

Strategic Product Manager

latest findings and trends in packet fronthaul

For the past decade, I had the privilege to work at Ericsson as a product manager bringing WDM optical fronthaul product portfolio and, recently, packet fronthaul portfolio to the market. A lot of learning over the years, and I continue to learn as we prepare and start deploying the first packet fronthaul networks.

Since the introduction of eCPRI specification in 2017 and in Ericsson RAN products a few years ago, Ericsson has seen an uptick in the interest in packet-based fronthaul solutions from operators all over the world. In this blog, I will give a quick update on the latest trends and findings in this area.

For a quick refresh, majority of the current RAN deployments have radio units on top of a tower or rooftop with the associated baseband processing units located in close proximity. The connectivity between the radios and baseband processing units is via a fiber-based CPRI or eCPRI interface. CPRI interface is TDM based and is implemented in many existing radios. eCPRI, on the other hand, is Ethernet based and currently mostly implemented in MIMO AAS (for more info please see Embark the evolution to Packet Fronthaul paper), and all new Ericsson radios will support eCPRI.

eCPRI, being an Ethernet based interface, makes it possible to use a packet network to transport eCPRI between radio and baseband processing units. This packet fronthaul transport network can facilitate the deployment of centralized RAN and Cloud RAN (see Cloud RAN blog), allowing RAN functions to be deployed in different physical locations in the network. In addition, the flexibility introduced by packet fronthaul network facilitates the ability to introduce automation to allow resource pooling and better life cycle management.

Now, let’s go straight into some of the trends and findings we have seen recently in packet fronthaul.

1. Radio over Ethernet is a niche solution

Currently, there are a couple of solutions in the industry to transport CPRI over the packet fronthaul network. One solution is Radio Over Ethernet (RoE), where CPRI are encapsulated over the transport network and regenerated on the other end. The other solution is a conversion of CPRI to eCPRI by distributing some of the baseband processing functions to the converter. What we have found is that dense radio configurations commonly found in urban macro sites make RoE-based solution mostly infeasible for Centralized RAN deployments due to the amount of transport bandwidth required. Using RoE, a 100GE interface does not have sufficient bandwidth to transport fronthaul traffic between a C-RAN hub site and its macro antenna site. In addition, RoE solution introduces additional latency, which limits the distance between a C-RAN hub and macro sites as RoE function is needed at the hub site. On the other hand, a solution that converts CPRI to eCPRI allows the same urban macro sites to use up to a single 100GE interface for transport between a C-RAN hub site and its macro antenna site with no latency overhead.

2. Proper synchronization solution is not just about Class C PTP T-BC

Radios depend on accurate sync information so they can transmit/receive signals without causing interference and support advanced RAN coordination features. In packet fronthaul, the synchronization information is distributed using Precision Time Protocol (PTP) and associated ITU-T telecom profiles. Router/switch vendors have over the years implemented PTP with enough accuracy (e.g., Class C), allowing them to be used in a packet fronthaul network. However, inaccuracy in synchronization distribution can be introduced from other areas. For example, asymmetry in transmit/receive fiber length of just a few meters may cause synchronization error greater than synchronization error introduced in router/switches. BiDi optics can be used to address this latency asymmetry issue and also reduce the number of fiber strands needed.

3. Increasing Cloud RAN/O-RAN activities are driving more interest toward packet fronthaul

With Cloud RAN/O-RAN activities at many operators, packet fronthaul naturally is a topic being heavily discussed. This is because Cloud RAN/O-RAN is about distributing RAN functions that used to exist in the Baseband processing unit across different platforms and different physical locations. The packet network is the glue that connects them all together so the solution can function as a system. We see that areas such as security and management are introducing new requirements into packet fronthaul networks.

using fronthaul gateways

In summary, while the journey of packet fronthaul has started, its use for centralized RAN and Cloud RAN deployments is still evolving. From a technology perspective, new RAN features and new devices will place new requirements on the network. Operationally, new methods and procedures to allow deployment, monitoring, and troubleshooting of these disaggregated network solutions are to be developed. Different operators have different strategies for evolving their networks. At Ericsson, we possess all the necessary pieces to support our customers in the path they choose, regardless if it is Distributed RAN, Centralized RAN, or Cloud RAN/O-RAN.

The Ericsson Blog

Like what you’re reading? Please sign up for email updates on your favorite topics.

Subscribe now

At the Ericsson Blog, we provide insight to make complex ideas on technology, innovation and business simple.