Home > Communication Service Provider Solutions > Enabling Telecom Transformation > RAN Pooling - The case for RAN pooling with Cloud/Virtualized RAN > Deployment and pooling flexibility with C-RAN
The disaggregated RAN, when containerized as cloud native applications, is no longer a monolith and the disaggregated NFs, which include the Centralized Unit (CU), Centralized Unit – User Plane (CU-UP), Centralized Unit – Control Plane (CU-CP), and Distributed Unit (DU), can be placed in different sites or locations, as per deployment needs and capabilities. RAN NFs can be pooled at different levels and locations of the deployment, again based on deployment needs and capabilities.
Figure 3. Deployment and pooling options with C-RAN
Figure 3 illustrates five C-RAN deployments with different pooling options [1]. Each has its advantages and tradeoffs.
Centralized RAN, which pools RAN NFs for many cell sites, offers maximum resource-sharing or pooling benefits in terms of Statistical Multiplexing Gains (SMG) and the best Total Cost of Ownership (TCO). SMG and TCO improvements are discussed later in this paper. Centralized RAN, however, requires an expensive high-performance and low-latency fiber fronthaul. Using a 7.2 Lower Layer Split (LLS) provides an excellent tradeoff on lower capacity and lower performance fronthaul. A Split RAN deployment offers maximum pooling of virtualized Centralized Units (vCUs), with the DU located at the cell site. Unlike Centralized RAN, Split RAN does not require large bandwidth and a performance-critical fronthaul. A Dual Split RAN has two levels of pooling, where virtualized Distributed Units (vDUs) at the aggregation sites are pooled to serve multiple cell sites, and a much more centralized pooled vCU covering multiple aggregation sites. Dual Split RAN is a compromise between a fully Centralized RAN (pooled vCU and vDU) and Split RAN (pooled vCU; no vDU pooling). Dual Split RAN obtains the benefits of vDU pooling, with much lower fronthaul capacity and performance requirements than Centralized RAN. Dual Split RAN deployments could be favorable in rural or suburban areas. A more Centralized RAN (with LLS 7.2 split) is favorable in urban areas due to more availability of fiber for fronthaul and much smaller and loaded cells for aggregation.
Depending on the use case they need to support, Communications Service Providers (CSPs) can dynamically and efficiently deploy one of these C-RAN NF pooling options. For example, to support a use case such as a local breakout or targeted public events with low latency requirements, the CSP may prefer a Dual Split RAN deployment, with pooled vDU, vCU-UP, and core network User Plane Function (UPF) at the aggregation sites or cell sites, and vCU-CP in a more centralized location pool. This is much harder to achieve with traditional D-RAN.