Welcome to the Network Intelligence ETI
The Network Intelligence (NI) sub-committee aims at supporting and endorsing research towards embedding Artificial Intelligence in future software-defined networks and programmable forwarding planes.
Future networks will need to embed intelligence towards agility, resiliency, faster customization, and security. Their advent represents an opportunity to embed intelligence while designing them, rethinking how we can solve networking problems using AI techniques.
The vision of embedding intelligence into the network will allow greater level of automation and adaptiveness, enabling
- Faster deployment (from months down to minutes)
- Dynamic provisioning, in line with the dynamic nature of network functions
- End-to-end orchestration, to ensure coherent deployment of IT and network infrastructures, and service chains
- High resiliency and availability of networks and services
Topics of Interest
The Network Intelligence (NI) sub-committee will cover numerous and multidisciplinary topics that are of importance to the ComSoc community, including:
- Declarative policies (“intents”) for orchestration and management: including Natural Language Processing and Understanding (NLPU) for service deployment, change, and assurance
- Learning techniques: supervised, unsupervised and reinforcement learning (forecasting, clustering, and classification techniques) for resilient networking
- Optimal resource allocation and placement and network action recommendations
- Knowledge base: graph database and advanced data mining techniques to ensure coherency of emerging networks (SDN, NFV, Programmable Forwarding Planes, Cloud, and 5G)
- Autonomic Management for Software-Defined Networks
- Self-configuration, self-optimization, self-healing, and self-protection in programmable and software-defined networks
- Self-optimization for dynamic controllers and virtual network functions placement
- Policy-based management, including imperative, declarative (intent), and other paradigms
- Learning and reasoning techniques for programmable networks
- Data analytics and machine learning for autonomic management
- Autonomic based service lifecycle management and orchestration
- Autonomic resource allocation and configuration in virtualized infrastructures
- Adaptive scheduling in cloud computing environments
The NI sub-committee will bring together (cross-fertilization) competences in network and competences in AI towards better, agile, and dynamic smart networks that become a must for the foreseen network transformation.