The DCCF coordinates the collection and distribution of data and prevents overlapping data subscriptions and notifications. The DCCF also forwards or retrieves data from the ADRF, which is a data aggregator that also handles data lifecycle management. The DCCF can also coordinate data collection/delivery via the MFAF, which formats and processes data received from the network functions. NWDAF MTLF uses Amazon SageMaker for machine learning (ML) lifecycle operations, including training and deploying ML model inference endpoints to be used by AnLFs for predictions. NWDAF uses Amazon Simple Storage Service (Amazon S3) backed AWS Lake Formation Data lake to store processed analytics, inference results, and trained models. NWDAF compares current data values against historical, expected, and predicted MoS (Mean Opinion Score) values to derive an Observed Service Experience analysis, which may be sent to the OAM for network optimization.
