Evolution of Artificial Intelligent Plane

TL;DR

This paper surveys the development and techniques of the AI plane in networks, providing a comprehensive framework for comparison, highlighting challenges, and guiding future research in network automation.

Contribution

It introduces the first complete survey and comparison framework for AI plane techniques across various network types, addressing key challenges and design considerations.

Findings

Provides a systematic comparison of AI plane techniques

Identifies key challenges and design considerations

Highlights future research directions in network AI

Abstract

With the growth of the internet, it is becoming hard to manage, configure and monitor networks. Recent trends to control and operate them is artificial intelligence based automation to minimize human intervention. Albeit this concept has been introduced since a decade with several different names, but the underlying goal remains the same, which is to make network intelligent enough to assemble, reassemble if configuration changes, and detect a problem on its own and fix it. As a result, in addition to Data Plane, Control Plane and Management Plane, a new plane called Artificial Intelligence (AI) Plane is introduced. Our main objective is to analyze all major AI plane techniques, frameworks and algorithms proposed in various types of networks. We propose a comprehensive and network independent framework to cover all aspects of AI plane, in particular we provide a systematically means of comparison. In conjunction to make AI plane understand simpler, this framework highlights relevant challenges and design considerations for future research. To the best of our knowledge this is the first survey report which represents a complete comparison of AI planes with their investigation issues in several types of networks.