The telecommunication industry is taking steps towards generating autonomous networks. Various standards organizations (e.g. 3GPP, ETSI, TM Forum) and open source organizations (ONAP, O-RAN) are defining mechanisms to implement management concepts such as Intent Based Management and closed-loops. So, as advances in this field accelerate, how do we get started with zero touch service operations?

Zero touch service operations is all about the automation of a series of processes along the service lifecycle, such as Service Design, Service Onboarding, Service Activation/Instantiation, and Service Maintenance and Operations. This automation can bring important benefits, as discussed in this blog ‘Zero Touch Service Assurance: A Tale of Two Approaches’.

However, if we take a closer look at these processes, they can also be defined from an autonomous management point of view. Both the 3GPP (TS 28.100) and TM Forum (IG1218) breakdown a process into different lifecycle functional categories and for each category there is a level (x) where machines take over from humans to automate that part of the lifecycle process.

zero touch operations
TM Forum – IG1218 Autonomous Networks – Business Requirements and Framework v2.0.0

In regards to the automation process within the industry, the Service Design, Service Onboarding, and Service Activation/Instantiation processes have seen significant progress in recent years – where Service Onboarding and Service Instantiation, to a large degree, are automated. In particular, Service Onboarding is mostly about automating the execution, and therefore is probably the most advanced one out of the three – from an automation perspective.

Let us now focus on the Service Maintenance and Operations processes. Naturally, this stage assumes that the services to be maintained are already deployed and active.

The maintenance processes are the first areas to apply full closed-loop due to these processes being the most time-intensive. While the previous service processes (such as Design or Activation) may be shorter in time, maintenance is always ongoing and is therefore a key area where the industry can gain value (ROI) by creating closed-loop automation.

Service Maintenance and Operations can be broken down into three key maintenance processes: SLA Preservation, Service Health, and Service Optimization. For these processes to be truly valuable to the industry, they need to achieve full closed-loop capability. To do so, we need to understand their level of automation readiness by assessing them across the entire cycle of Awareness, Analysis, Decision Making, and Execution.

The below table estimates the complexity of each automation category in the context of the key maintenance processes. The more complex processes require AI/ML to support them. Since the complexity depends on the specific use cases, it is possible that use cases will be in different categories of complexity.

Process SLA Preservation (Proactive & Reactive) Service Health (Healing) Service Optimization
Awareness Standard Moderate Moderate
Analysis Standard-Moderate Moderate-High High
Decision Standard-Moderate High Moderate-High
Execution Standard Standard Moderate


Service Assurance and Analytics is about Awareness and Analysis. However, for successful, complete, fully automated closed-loops, we have to consider the complexity of all the closed-loop stages. If we look at the three processes, we can observe the following:

  • SLA preservation: the relatively more complex part may be in the Analysis and Decision stages. Yet, in comparison to the other two processes, these two are less complex.
  • Service Health: In general, it requires a higher complexity of Analysis than SLA Preservation. However, it includes a very wide range of use cases that require Analysis at a Moderate to High level of complexity. Some use cases, such as identifying root cause at a physical layer (non-functional cards, cable cuts, etc.) are on the Moderate side, while other use cases, such as identifying non-optimized configurations, may be much more complex. The key reason for this is that many of today’s systems are data systems and not yet knowledge systems, so for the high-end use cases they sometimes find the “Root Cause Symptom”, which is the closest to the reason but is not necessarily the “Root Cause Reason”. Service Health also requires good Service Impact analysis based on various data types (e.g. alarms, KPIs, usage records) – which is not easy to achieve.
  • Service Optimization: typically requires AI/ML algorithms for analysis with a moderate to high complexity decision making process.

So how do we get started with zero touch service operations?

I would recommend SLA preservation as the first area to start applying closed-loop automation. The next candidates for closed-loop automation are the Service Health use cases where the complexity of the analysis is ‘Moderate’ and you also have the Root Cause Analysis algorithms that can deliver the “Root Cause Reason”.  Service Optimization, however, may require the highest level of Analytics and the specific algorithms to support the different use cases. There are many kinds of possible optimizations and a lot of them will require their own specific algorithm. Yet, once the right algorithm for a use case is deployed, the rest of the closed-loop stages may be more easily achievable.

For more information on how TEOCO can help you realize your zero-touch service assurance automation goals, visit