Objectives
RAN architecture design
The design of an open and modular 6G RAN architecture, which improves energy efficiency and lowers operating costs.
Workload acceleration
The optimization of key RAN functions to be accelerator-friendly, and able to be executed across a wide range of devices.
RAN Intelligence across different architectures
The execution of RAN intelligence on heterogeneous hardware, including emerging compute platforms.
AI orchestration system
The implementation of an AI orchestration system, which will dynamically enforce policies based on operator intents, while also ensuring execution security.
Lifecycle management through AI pipeline
The development of an AI Pipeline, which is hardware-aware and collects telemetry to automate the lifecycle of models used in the RAN.
Verification
The validation of SHARC’s innovations by a variety of specific use cases.