The operator interest in open RAN solutions, and willingness to invest in trials, will open new doors for a variety of smaller vendors which have been showing interesting solutions over the years, but have been confined to a small cell market which grew more slowly than most had expected. One such supplier is Belgium’s Accelleran, which has been developing a dyamically reconfigurable RAN platform based on ORAN interfaces.
Accelleran’s vRAN architecture centers on a virtualized controller called dRAX Open Interface RAN Intelligence, which supports network functions in a cloud-native, microservices-based environment. The leap from a virtualized to a fully cloud-native environment, with all the flexibility and reconfigurability that brings, is still in the future for most operators, but at least the low level of vRAN adoption means that they will find it easier, when the time comes, to go straight to a cloud-native RAN, rather than undertaking a complex migration from a virtual machine architecture, as some face in the core.
Although most operators believe the cloud-native RAN is a few years away, as they wait for standards to stabilize and architectures to mature, there is shorter term activity in the small cell arena, where companies like Accelleran have focused. Centralized, sometimes virtualized, clusters of small cells are an increasingly popular topology for enterprise networks, and this approach is expected to grow rapidly from 2020, along with 4G/5G densification and the need to achieve full indoor coverage at last.
The dRAX system supports Accelleran or third party radio units. It uses open, standard interfaces to connect the RAN controller and the cells’ control planes within dRAX, and between the controller and the RAN user plane functions and the edge compute applications. It is cloud-native and can work with virtual machines (OpenStack) or containers (Kubernetes). The containers communicate over a common messaging bus.
This allows the centralized functions in a vRAN to run on embedded SoCs as well as on cloud or enterprise servers, and a virtual network function can be dynamically redeployed between these environments, and service-chained with edge compute functions.
The control and user plane location are independent and flexible, and user plane processing can be independently assigned to each UE and service bearer. Open APIs allow any radio head or small cell to be integrated into a dRAX network, and the controller can support a single microserver at the edge to a cluster of hundreds of cells.