Industry Specification Group (ISG) Experiential Networked Intelligence (ENI) Activity Report 2023

Chair: Raymond Forbes, Huawei Technologies

Responsible for developing standards that leverage Artificial Intelligence (AI) mechanisms for improving the operator experience to more quickly recognize and incorporate new and changed knowledge, and hence make more rapid actionable decisions as well as assisting in the management and orchestration of the network.

Technologies such as Software Defined Networking (SDN), Network Functions Virtualization (NFV) and network slicing are making networks more flexible and powerful. However these same innovations are also making networks harder to manage efficiently. 

The use of Artificial Intelligence (AI) techniques in the network supervisory and management system can help address some of the challenges of future network deployment and operation.

ETSI’s Industry Specification Group on Experiential Networked Intelligence (ISG ENI) develops standards that use AI mechanisms to assist in the management and orchestration of the network. The group’s work focuses on improving the operator experience, using closed-loop AI mechanisms based on context-aware, metadata-driven policies. This enables the ENI system to recognize and incorporate new and changed knowledge, and hence make actionable decisions.

ISG ENI has specified a set of use cases and the derived requirements for a generic technology independent architecture of a network supervisory assistant system based on the ‘observe-orient-decide-act’ control loop model. This model gives recommendations to decision-making systems, such as network control and management systems, to adjust services and resources offered based on changes in user needs, environmental conditions and business goals.

By developing standards that use AI mechanisms to recognize new or changed knowledge – and thus make actionable decisions for operators – the work of ISG ENI enables an efficient intelligence-based assistance to SDN and NFV deployments which will in turn manage and orchestrate the network.

The group accordingly considers a range of ENI related issues including architecture, AI, security and Proof of Concept (PoC) frameworks, while coordinating activities with other entities within ETSI and externally.

Published in January 2023, ETSI White Paper No.51 describes the design of a novel cognitive network. “ENI Vision: Understanding the Operator Experience Using Cognitive Management” explains how ETSI’s ENI system architecture intelligently manages, predicts, adjusts and optimizes network behaviour using cognition management, thereby enhancing the operator experience. By modelling human decision-making processes to better comprehend the relevance and meaning of ingested data, the ENI system can experientially learn to improve its operation and performance. Key takeaways include:

  • Exploration of the ENI cognition model
  • Examination of multiple facets of the cognition process, including context awareness, situation awareness and the formal cognition process (perception, comprehension, and action)
  • Similarity between human cognition and ENI cognition, and the use of reactive, deliberative and reflective processes to implement ENI cognition

The paper also describes relevant technical details of the ENI System, as well as a number of recent innovative Proof of Concept projects. 

During 2023 the continuing focus of the group’s technical work was on ENI Release 3 and Release 4.

Published in December 2023, GR ENI 035 V4.1.1 ‘Definition of IP networks autonomicity level’ defines network autonomicity features and levels for IP networks, including the intelligent characteristics at each layer (from Level 0 to Level 5) and closed-loop management process. With the report specifically defining the evaluation criteria for the determination of the autonomicity level in the domain of the IP network, it has been possible to obtain measurable criteria to score actual benchmarks. This work is foreseen to be expanded to other domains during 2024, potentially involving domains such as Datacentre.

By the end of 2023, a number of further Release 4 deliverables were declared ‘stable’ – or approaching stability – in anticipation of publication in early 2024:

  • GR ENI 010 version 2 - Evaluation of categories for AI application to Networks
  • GR ENI 017 version 2 - Overview of Prominent Control Loop Architectures
  • GS ENI 030 early release 4 - Transformer Architecture for Policy Translation
  • GR ENI 031 - Construction and application of fault maintenance network knowledge graphs
  • GR ENI 015 - Processing and Management of Intent Policy
  • GR ENI 032 In-situ Flow Information Telemetry (IFIT) Deployment Scenarios

Several new Work Items in the scope of AI transformer ‘large language models’ are currently in drafting.

Meanwhile completion of a number of further Work Items related to Release 3 was achieved during the year, including updates to existing Release 2 deliverables. A number of these deliverables were accordingly published during 2023, notably:

See the full list of ISG ENI Work Items currently in development here.

Developed according to our Proof of Concept Framework, ENI Proofs of Concept (PoC) are intended to demonstrate ENI as a viable technology. Results are fed back to the relevant Industry Specification Group for each PoC. A list and descriptions of all current ENI PoC are available on the public ETSI ENI wiki pages.

During 2023 ISG ENI maintained its active liaison with other ETSI groups, namely: ISG NFV, ISG ZSM, MEF, ISG F5G, 3GPP TSG SA5, ITU-T SG13, SG12, SG11 and ITU-T FGML5G.

The group also continued to interact with other ETSI entities, including TC CYBER and ISG SAI, to ensure that all communications between ENI and Assisted Systems operate securely. In addition, it continued to coordinate activities closely with OCG AN on Autonomous Networks and OCG AI on the AI Act SR, and with related ISGs and TCs.

For more information on current ISG ENI activities read the Chair’s regularly updated blog here.