5G is becoming more and more widely adopted by both mobile networks and private network operators. To meet the demands of new services and applications, 5G needs to be able to not just scale at pace but become a more flexible offering, able to be implemented in a number of different scenarios, and naturally incorporate new innovation such as virtualisation.
The shift towards network virtualisation has been accelerating in a wide range of technological sectors, where the use of cloud services and open interfaces is commonplace and well understood. However, in mobile networks, this doesn’t generally hold true and it is still common to find proprietary interfaces and single vendor solutions. This is changing however. Today, the 5G core is commonly cloud-based, boasting benefits such as improved scalability and lower operational costs. A system like this can very easily be set up — you configure an instance and it’s good to go.
However, the RAN is much more complex and comes with a number of challenges when trying to virtualise the systems that make it up.
What is ORAN?
A groundbreaking initiative is enabling new ways to roll out 5G mobile networks, all the way from the radio front end through to the core network, known as Open RAN (ORAN). The ethos behind ORAN is the opening up of previously proprietary, horizontal architectures with open, software-based ones that are able to run on commercial servers.
The resulting key benefit is that 5G network deployments are opened to new entrants by lowering the barriers to entry, enabling new partnerships, innovative products, and better service offerings to customers. This in turn benefits those deploying 5G networks by allowing a much richer ecosystem of vendors, and a number of different network deployment models to be envisioned in a cost-effective manner.
Virtualising the RAN
Virtualisation enables systems and networks to run on commercial off the shelf servers (COTS) in the cloud or on the edge. These are much more cost-effective to implement and easier to run and maintain as users can pick who they want for their software, hardware and radio interfaces, rather than being tied into a single vendor.
However, RAN architecture can be very complex, which is why proprietary systems from single vendors have been the deployment option of choice to date. ORAN enables various elements of the RAN to be split into functional units, which can be provided by different vendors, centralised where needed, with resources managed more effectively as you get closer to the edge of the network. Three key components form the RAN – the RU, DU & CU (Radio Unit, Distributed Unit & Centralised Unit), supported by a RIC (RAN Intelligent Controller). This can then be linked to existing management and orchestration systems, as well as existing 5G cores so that the network is run in the same way as now.
One can imagine an architecture with many customer facing cells, where the radio system is close to the customer, but the management layers are hosted centrally to optimise efficiencies and minimise the amount of physical equipment needed to operate a RAN. These spread out deployments naturally create physical splits in the architecture. In practice, these ‘splits’ between the different components need to have high bandwidth, low latency connections to ensure that the quality of service can be maintained and that multiple units can be managed in unison. Connecting these elements wirelessly with fronthaul and backhaul so that maximum deployment flexibility can be met is where high-bandwidth mmWave networks can play an important role.
The current ORAN landscape
A number of ORAN deployments, projects and testbeds have been coming to life designed to promote interoperability of different vendors’ equipment and accelerate the adoption of ORAN technologies in mobile networks.
It is still early days for ORAN deployments, but creating an environment where innovation is possible is crucial for the success of ORAN. Only by working together can interoperability be achieved. The success of these early deployments will really dictate the role of ORAN in mobile networks for years to come. Achieving commonality of features and performance across a multi-vendor infrastructure is not an easy task, but by opening up the market to proven models, such as cloud platforms, many issues around complex integration tasks can be resolved or removed from view. These open models have worked in a number of different sectors, and RAN deployments should be no different.
There are a number of projects which are looking at how to deploy ORAN most effectively. One such example is the Liverpool 5G Create project that uses mmWave backhaul to operate a small cell network with roaming capability. Trials and projects such as these are demonstrating how ORAN architecture can provide a strong blueprint for consistently high bandwidth and low latency, strong network performance and enhanced customer experience.
Barriers to adoption
Network virtualisation has been proven to provide greater cost efficiency and flexibility to network operators. However, the ORAN initiative has yet to demonstrate in practice that it can strike the right balance between performance and resilience.
One of the main challenges is that pilot projects are being led by smaller vendors rather than large players. This means that ORAN deployments may well find success in certain types of network deployment such as small cell or private networks, rather than large scale macro networks. Another key challenge is security. By opening up the interfaces between the various parts of the network, the attack surface is expanded and new visibility gaps can arise, creating new cybersecurity risks for operators. All of this has to be managed in a network that has to scale up to support a wide range of services and applications whilst maintaining network uniformity, reliability and agility.
Also, the industry as a whole faces challenges in defining a level playing field and clarify rules of engagement, moving from old deployment models to new open systems that create an environment that supports competition between suppliers.
Why 5G private networks are set to drive ORAN adoption
5G private networks are an attractive model for ORAN for both industries and operators than can build them. ORAN implementations within private networks are still very rare. Businesses that have implemented private networks to date have normally used single vendor technology. Having a multi-vendor open environment available to them would mean that the barriers to adoption of 5G private networks would be lowered considerably. Not only would the skills required to implement a 5G network be more widely available, the infrastructure overhead would be reduced considerably, allowing COTS hardware to be deployed in place of expensive, proprietary equipment.
It is likely that mobile operators will also look to private networks as a good way to prove the viability of ORAN. Typically, private networks are much smaller, less complex and have less services running over them than public macro networks. By implementing ORAN-based private networks, operators will be able to prove the ORAN model decisively and iron out any issues before scaling into larger national networks.
Open networks have the opportunity to drive growth in the private mobile network market, with operators, new entrants, hyperscalers and systems integrators able to build solutions that are much more flexible and resilient to change than they are today. Although this model is still a way off, the current ORAN testbeds are providing a growing business case for organisations looking to design and implement private networks.
