Survey of Intent-Based Networks and a Methodology Based on Machine Learning and Natural Language Processing

Intention-based networks (IBN) have been developed in order to solve the problems of traditional networks such as: inefficient performance management, security and applications, the company spends too many resources on troubleshooting, instead of investing in new business services. These problems are solved with IBN because this technology combines artificial intelligence and machine learning, in order to transform the commercial intention of users into strategies for configuration, operation and maintenance of the network. IBN is an architecture in which Cisco has invested in recent years; however, it is a broad architecture to be explored and adopted by organizations. Because of this, in this research, a survey is developed using a methodology and algorithms in order to examines the elements for implementation, applications in different environments, tools, protocols, algorithms and trends of intent-based networks. For the development of the survey, two processes have been used: the first one consists of a general method which is used for text extraction, natural language processing of scientific articles and exploratory analysis. This method allows us to obtain the most relevant topics within the data set. The second process starts from the processed data set and then applies two filtering methods. The first method which we have named “Calculation of Importance-Based Approach” and the second method which we have named “Bag of Words Extraction by Cosine Similarity-Based Approach.” These methods help us to obtain the items most related to the topics obtained in the general method to develop the survey of intention-based networks. These processes may be useful in other fields of interest.