Easily Designing and Deploying AI enabled Network and Telecom services

Currently, there do not exist any open-source systems that offer streamlined and easy development and productionalization of AI/ML enabled Intelligent Network and Telecom services. This is a big problem the Network operators/Service providers face today even though Network Function Virtualization (NFV) and Software Defined Networking (SDN) have become ubiquitous and the Next Generation networks aim to be more softwarized and AI-driven. This paper aims to show how to quickly and easily design Network, Telecom and Network Security services and deploy them into a containerized cloud environment. This paper also shows how to make the Network/Telecom services AI/ML enabled and deploy them in the same way into the containerized cloud environment. We will show using ONAP (an open-source whitebox Network Operating System and management platform) and Acumos (open-source AI platform for creation, sharing and management of AI models and solutions), and specifically using their DCAEMOD and Acumos Adapter components, that this is very easy to do. To serve as an example, we will create a Network service to analyze and classify malicious network traffic. Using the same framework and tools, we will deploy the created service as a set of communicating, containerized microservices into a Network Operating system and Management platform like ONAP which is cloud-agnostic and can be run on any cloud. The same created service can also be deployed directly into any other OSS or on any other cloud platform.