Automation of Network Configuration Generation using Large Language Models

The life cycle for a service provider (SP) to launch a new service or tariff plan often requires months of planning and testing. The SP has traditionally used OSS (Operation support systems) and BSS (Business support systems) which are large, non-standard systems providing life cycle management related to launching new digital services (e.g. internet, voice, SMS, IoT) and tariff plans. These OSS/BSS systems, being multi-vendor and custom implementations, involve significant costs and a timeline of months to a year to roll out the service. This problem worsens with evolving technologies such as 5G. This paper addresses the specific need for faster provisioning of 5G network services and their corresponding tariff/billing plans, thus shortening the duration for launch. To provision the network, a combination of a deep neural network and a large language model (LLM) is proposed in this work to automate the generation of network configurations.