Artificial Intelligence (AI) in Test Systems, Testing AI Models and the ETSI GANA Model's Cognitive Decision Elements (DEs) via a Generic Test Framework for Testing ETSI GANA Multi-Layer Autonomics & their AI Algorithms for Closed-Loop Network Automation

This Technical White Paper No.5 is about Demo-4 (of a series of Demos planned for 2018/2019 and beyond) of the ETSI PoC (Proof-Of-Concept) on 5G Network Slices Creation, Autonomic & Cognitive Management and E2E Orchestration; with Closed-Loop (Autonomic) Service Assurance for Network Slices; using the Smart Insurance IoT Use Case.

The Technical White Paper covers the following key topics in the target of operationalizing the ETSI GANA Model Standard (ETSI TS 103 195-2) in the context of 5G Networks and reveals Gaps in Standards that should be addressed by the industry while implementing GANA autonomics (control-loops) in network architectures and associated management and control architectures. The following aspects are covered in the results of the PoC:

• The Newly Launched Work Item (WI) in ETSI TC INT on "AI in Test Systems and Testing AI Models", and the call for contributions to WI
• The Benefits AI brings to Test Systems (embedding AI in a Test System/Component)
• Testing AI Models for autonomic and cognitive management & control of network resources, parameters and services
• ETSI GANA as Multi-Layer AI Reference Model for Implementing Autonomic Management & Control (AMC) of Networks and Services (including 5G Network Slices)
• Stakeholders that should play certain roles in the context of Test and Certification of GANA Cognitive DEs and GANA Knowledge Planes for AMC
• General Approach to Designing the Test Systems/Components for Testing Autonomics AI Models, and Challenges in AI Models’ ability to cope with 5G Network Dynamics that need to be taken into consideration
• Illustrations of Types of Standardizable Metrics that should be target for Measurements and Assessments in Testing and Certification of AI Models of Autonomic Components/Systems
• Capabilities of Softwell Performance AB’s AI-empowered Performance Test Solutions that held address Challenges in Performance Testing of 5G Networks, Products and Slice Services
• Capabilities of Rohde & Schwarz Test Solutions that play a role in Enabling Autonomic Management & Control of 5G Slices
• Capabilities of DATAKOM Solutions that play a role in Testing AI Models for Autonomic Management & Control of 5G Slices
• Generic Test Framework for Testing ETSI GANA Model’s Multi-Layer Autonomics & AI Algorithms for Closed-Loop Network Automation