Autonomous Networks (AN) have gained a lot of attention, but the term is often misunderstood, leading to numerous myths and misconceptions. As one of the earliest AN practitioners, Infosys Consulting has been partnering with global telecom giants to craft tailored AN implementation roadmaps. In this blog, written by a senior AN expert, we share our early experiences with AN, and provide insights on C-Level questions about Autonomous Networks.

Telecom circles have been buzzing with “Autonomous Network” talk in 2024 — catching plenty of C-suite eyeballs with the perceived big promise of the term. But there is a ton of confusion about what Autonomous Networks actually are.

At Infosys Consulting, we’ve been partnering with telecom players worldwide to create tailored AN technology, process, and operating model transformation roadmaps. When we sit down with CTOs and network leaders, everyone has their own take on what an Autonomous Network actually is. Dig a little deeper, and you’ll find plenty of skepticism about whether Autonomous Networks can deliver on their hype.

Let’s set the record straight.

To start with, what are Autonomous Networks? Autonomous networks are essentially intelligent networks. According to TM Forum’s Autonomous Network Levels Evaluation Methodology Whitepaper  IG1392, Autonomous Network maturity can be categorized into six levels (0 to 5):

• Level 0: Fully manual operations
• Level 5: Fully autonomous network with intelligent closed-loop capabilities.

As networks progress through these AN maturity levels, human intervention decreases and intelligence across the AN’s closed-loop capabilities increases—Awareness, Analysis, Decision, Execution, and Intent/Experience.

One of the primary misconceptions around ANs revolves around network automation, with many professionals mistaking the automation of networks for autonomous networks.

Let’s be clear – network automation is not an Autonomous Network! Communication Service Providers (CSPs) have been using the TM Forum industry definition to separate true Autonomous Networks from typical Network Automation. To demystify the concept further, Infosys Consulting decomposed AN into basic building blocks to explain interconnected and interchangeable terms of references. While these capabilities are interconnected with AN, they are not synonymous.

Here is a breakdown of the terms with brief descriptions to cut through the jargon:

• Network Observability: Monitors and explains the network’s real-time operating state.
• Network AIOps: Detects anomalies, diagnoses cause, and suggests possible solutions.
• Closed Loop: Automates entire cycle of anomaly detection, cause analysis, options evaluation, and resolution; utilizing Observability, AIOps, and self-healing Orchestration.
• Autonomous Network: Operates Closed Loops autonomously to ensure network service resiliency and reliability.
• Intent-based Networks: Fully autonomous in managing service intents using Autonomous Networks. Intents are goals defined with context and constraints.

Infosys Consulting teams have garnered expertise in ANs through early engagements with global telecom providers, as well with our contributions to industry fora.

Infosys Consulting: Pioneers with AN Evolution

Infosys Consulting pioneered one of the earliest AN maturity assessments for an Australian Tier 1 telecom player, using the TM Forum-defined AN level evaluation framework. Our efforts culminated in a comprehensive three-year technology roadmap, featuring over 50 initiatives aimed at achieving TMF AN Level 4 (High Autonomous Level) across various telecom network technology domains including 5G, Cloud, Transport, and IP.

Building on this success, we embraced the challenge with another Tier 1 telecom enterprise in Singapore. We leveraged Autonomous Networks fundamentals to transform their Network Operation Centre (NOC) processes, OSS and E/NMS technologies, and operating model to realize their vision of Autonomous Operation.

Our insights into Autonomous Networks and Operations, gathered through meaningful conversations and engagements with telecom clients, inspired us to contribute to industry whitepapers such as TMF IG 1392.

Infosys Consulting aims to bring clarity to real-life scenarios and challenges and to guide telecom companies with their AN implementation roadmap. During these engagements, our senior consultants have guided CXOs and helped them answer some crucial questions to aid their decision making.

Top 3 Questions on ANs from C-suite Leaders

These are the most prevalent questions in our interactions with CXOs of various telecom players, with our responses.

1.     Is TMF Level 5 Maturity (Fully Autonomous Network) vision achievable?

The journey toward fully Autonomous Networks (AN) at TMF Level 5 is both ambitious and complex. According to the TMF AN Survey, only 4% of telecom operators believe they will reach AN Level 5 by 2028.

However, with the rise of AI/ML, Generative AI (GenAI/LLM), and Cloud-Native architectures, certain limited-use cases for TMF AN Level 5 are definitely becoming viable, particularly in domains such as 5G and beyond. Several core technologies play a pivotal role here, such as:

• Containerized Network Functions (CNFs) that enable self-healing capabilities CNFs are advancing but are primarily deployed in mobile networks.
• Generative AI (GenAI/LLM) which facilitates intent translation and automated remediation by using design and run-time Network Knowledge. We should note that GenAI/LLM adoption remains in the Proof of Concept (PoC) and pilot phase,
• Network Ontology and Digital Twin are technologies that provide knowledge about network design and network live operating state, which can used for analyzing and simulating anomalies and their root causes and getting suggested corrective actions.  However, Network Ontology and Digital Twin are still conceptual, posing a barrier to achieving fully autonomous network operation.

2.     How does a Leader Build a Business Case for Autonomous Networks?

A compelling business case for Autonomous Networks must address three primary financial and operational factors:

• Experience – Revenue generated by improving the reliability and resilience of network experiences
• Economic – Cost savings achieved through autonomous operations across a wide range of use cases and scenarios
• Evolution – Technology debt mitigated by staying aligned with ongoing technological advancements

Together, these three pillars provide a foundation to translate AN maturity into measurable business value. To build this case, the following steps are key:

1. Identify the desired network operations objectives.
2. Link SMART (Specific, Measurable, Achievable, Realistic, Time-bound) Key Effective Indicators (KEIs) to these objectives.
3. Map AN technology capabilities that can directly enhance these KEIs.
4. Calculate the cost savings ($) per unit improvement in the identified KEIs.
5. Estimate the investment ($) required for implementing the necessary AN technologies.

This approach works particularly well for specific AN technologies, such as Network Observability and Anomaly Detection. However, justifying investments in foundational AN technologies—like Network and Service Inventory, Network Ontology, Orchestration, and Knowledge Graphs—can be more challenging.

3.     Will Network Operations Centers (NOCs) become Obsolete with Autonomous Networks?

While Autonomous Networks will optimize and simplify NOC operations, they will not eliminate them. In actual fact, the NOC’s role and operating model will evolve as automation and AI-driven intelligence advance. Here is how different tiers of NOC operations might be handled:

1. NOC Tier 0 and 1 tasks – Monitoring, incident logging, enrichment, and tracking will be largely automated for most scenarios due to advancements in Network Observability and Network AIOps (anomaly detection).

2. NOC Tier 2 and 3 tasks – More specialized functions, including deep diagnosis and complex problem-solving, will be streamlined through emerging technologies like Network Ontology, Digital Twin, Knowledge Graphs, and Generative AI (GenAI/LLM). These technologies will reduce the expertise and manual effort required for intricate operational tasks.

3. NOC Tier 4 operations – Functions that involve liaising with network vendors (partner ecosystems) for remediation will be simplified using domain and technology-specific GenAI/SLM (Small Language Models).

In a nutshell, NOC Tier 0 and 1 tasks will see a significant reduction in volume, with the majority shifting to closed-loop automation. NOC Tier 2, 3, and 4 operations, on the other hand, will be simplified with greater contextual insights, leading to faster diagnostics and resolution.

Key AN Insights for Telecom CXOs

Achieving AN Level 5

• The biggest barrier to achieving AN Level 5 is technology maturity. Physical infrastructure components, such as microwave antennas and optical fibers, also pose significant challenges that prevent full automation across all scenarios.
• The transition to Autonomous Networks will be gradual but inevitable, as AI capabilities mature, and cloud-native architectures unlock new possibilities.

Building a business case for ANs

• ·To strengthen the business case for AN, initiatives should incorporate a sufficient number of use cases to justify investments in foundational technologies.
• By spreading costs across multiple applications, these investments become more compelling.
• Initiatives with fewer use cases often struggle to secure funding, as the perceived return on investment may not seem justifiable.

The future of NOCs

• The shift toward full autonomy raises important questions about the future of traditional roles and structures within network operation and management.

Achieving AN Level 4 maturity and progressing towards Level 5 requires a structured approach to emerging technology adoption, process and operating changes, and business objective alignment on this journey.