Home > Communication Service Provider Solutions > Enabling Telecom Transformation > 5G Disaggregated RAN Performance Benchmarking Insights > Look-aside L1 acceleration
In the disaggregated RAN environment, DU usually implements High PHY Low Level Split 7.x aka L1 Acceleration. The “Look-aside L1 acceleration” architecture offloads some selected functions of L1, while the rest are processed in the CPUs by assigning the dedicated CPU cores.
As depicted in the diagram above for “Look-aside”, the CPU processes the UE traffic that arrives at the DU from the higher layers (PDCP through the Midhaul) using DU-L2 (RLC and MAC Scheduler). It then forwards it to L1 using FAPI P5 and P7 interfaces.
In this example, L1 offloads some of its tasks related to FEC base-band encoding, rate matching and scrambling of UEs data. After these tasks are completed, the CPU continues to run the other functions and encapsulate the IQ and control data into the eCPRI packets before sending to fronthaul. The same applies in the uplink direction—FEC decoding is offloaded from CPU before packets are delivered to the upper layers.