Earlier this month we wrote an analysis looking at network and service automation and why it was an increasingly important topic, especially in a 5G context.

But, while we captured the main automation market drivers, we left out one important consideration: the role of intent-driven networking and the importance of collaboration in making it possible.

The analysis is worth a quick read but, to save you some time, I can recap the main points. They are fairly straightforward.

With operators scaling their 5G networks and services in 2020, it’s fair to say we are now firmly in a period of commercialised 5G. Nearly 150 operators around the world have launched services, and affordable smartphones can put services in the hands of an increasing number of consumers.

But 5G comes with its own challenges (costs) for operators:

  • Network operations. 5G involves new radio access and core assets deployed in new architectures and places, all alongside legacy networks.
  • B2B operations. The enterprise sector is a major source of 5G optimism: for most operators, it represents their greatest hope for new revenue. Yet, much like 5G itself, moving into the enterprise comes with new network architectures and assets to be deployed, adding even more complexity into the picture.
  • Quality demands. Selling consumers or enterprises on the 5G promise will only be possible if services are reliable and deliver as promised. Against the backdrop of new network complexity and operations challenges, that cannot be taken for granted.

Now, recognising opex is the major cost centre for most operators, easily outpacing capex by up to four-times, it’s clear solving these challenges by simply throwing more resources or labour at them isn’t a palatable option. Network and service automation, then, takes on a renewed importance but through the lens of a journey where automated systems and processes can be put in place in a step-wise manner.

What is intent-driven networking?
With the background context out of the way, it’s time to introduce a new concept into the discussion: intent.

Intent Management. Intent-driven networks. Intent-based networking. The concept goes by many names, but the basic idea is a relatively simple one, if often accompanied by lots of technical detail.

Intent-driven networking aims to strip out the complexities associated with network policies (creating them, managing them, enforcing them in line with general business objectives) to limit the need for (and errors caused by) human capital. Of course, AI and ML play a major role in enabling this.

Now, if this seems like a very broad description of automation in general, you’d be correct. But the difference here is all about the term “business objectives.”

The notion of intent is built around network configurations and commands that are driven by business objectives. This means policies need to be user/business friendly and outcome-based. At the same time, it means business intent needs to be translated into specific actions in terms of resource allocation, policy enforcement and network/service monitoring, all with AI/ML tools doing their work in the background. When it’s in place, intent-management promises network administrators the ability to define a business outcome/intent with the networks AI/ML capabilities sorting out how to make it happen and then actually make it happen.

Can you have intent without collaboration?
On paper, then, the concept of intent is straightforward: a focus on business outcomes instead of network configurations aligns completely with the goal of reduced complexity and improved service quality.

But is implementation just as straightforward?

You probably already know the answer. It’s going to be a complex effort because operator networks, themselves, are multi-layered and complex. This may seem obvious, but it is an important reminder. A given operator will maintain many multiple OSS systems, as well as myriad element management (EMS) and network management (NMS) assets. Going forward, these will need to consider many new network infrastructure locations (think edge computing), from which new services will be coordinated and/or delivered. In most cases, the network will be built from a number of different vendors. And this is all taking place against the backdrop of a universe of operators, many of whom may try to drive intent-based automation forward in their own ways.

This comes with a very clear implication: we will need collaboration across many multiple dimensions.

Of course, strategically we will need collaboration across vendors and operators to ensure we can identify (infer) the intent of various applications and that APIs remain open enough to support all of this. More fundamentally, however we will need collaboration across OSS, EMS, and NMS solutions. Without this, the end-to-end view needed to deliver on automated intent management just won’t work. And, remember those new network locations? They obviously need to be a part of this equation, but maybe not in the way you’re thinking. Yes, these assets will need to be managed. But, as automated (intent-driven) service delivery is rolled out, operators will need to consider the best locations from which to have decisions made, closer to the core for cross-domain, global, non real-time decisions versus closer to the user where real-time performance requirements take precedence. This too will need close collaboration.

If the value of intent as a core component of network and service automation is so obvious, you might be asking why we didn’t address it in our analysis.

It’s a good question, with a simple answer.

While we know that operators are prioritising automation, it’s still unclear how they view intent as a part of those strategies. To be fair, vendors claim no shortage of operator collaborations around automation and even note intent as part of that. But, consider your average network slicing conversation. How integral is the concept of intent? It should be central to the very existence of slicing, but they aren’t always mentioned in the same breath.

Now, we can’t necessarily expect two emerging technologies, driven by different sets of interests, to have their terminologies align perfectly. Yet, this does highlight a disconnect and why we need to understand better how operators are thinking about intent.

While we looked at automation extensively in our last Network Transformation survey, we did not dig into the topic of intent. As we launch an update to our survey, look for further insights on the topic in the future.

– Peter Jarich – head of GSMA Intelligence

The editorial views expressed in this article are solely those of the author and will not necessarily reflect the views of the GSMA, its Members or Associate Members.