ONAP Based Pro-Active Access Discovery and Selection for 5G Networks

This paper enhances the functionality of analytics and policy framework of ONAP to solve the problem of access discovery and selection of radio access networks, which is currently manually configured by the operators. We propose to automate the policy creation and handle it dynamically, based on current and historic data collected from various network nodes based on three main services (open to accomodate new services) of 5G i.e. Enhanced mobile broadband (eMBB), Ultra-reliable Low Latency communication (URLLC) and Massive Machine type communication (mMTC). Learning, analysis and prediction of the real and non-real time data can help to form dynamic access discovery and selection policies on the go with help of ONAP's DCAE (Data Collection, Analytics, and Events) and policy framework. To verify the effectiveness of our proposal we created a test bed with different access points and blended SOC with ONAP. Based on three different use cases we collected real time and non-real time traffic traces for our SOC to form dynamic policies with the help of DCAE and policy framework for selection of best available gNBs or eNBs. We compare our proposal with legacy networks and existing research works in the literature.