Transport automation: Tackling scale and complexity

Network complexity has gradually increased in the past few years, becoming as big a challenge as handling capacity demands. For many operators, reducing complexity through automation is now a priority. They are pursuing emerging technologies to quickly deploy network automation and make it as efficient and accurate as possible.

This blog discusses recent innovations around intent-based networking, artificial intelligence/ machine learning (AI/ML) and active assurance — and their interrelation. It is based on key findings from Heavy Reading's third annual Open, Automated, and Programmable Transport Networks global operator survey.

Intent-based networking

In Heavy Reading's survey of 80 operators worldwide, we asked respondents to weigh in on a set of automation-related statements by indicating degrees of agreement or disagreement with each. At the high end, 43% of operators surveyed said they "strongly agree" that increased network complexity is a key reason to invest in more automation. Additionally, 37% "strongly agree" that they need to become more intent-driven, and 35% "strongly agree" that more automation is needed to cope with increasing scale and automation demands.

These statements point to increasing network complexity and scale driving the need for advanced automation technologies that can simplify network operations by automating more of the requisite configurations for a given service with a model-based approach. Intent-based networking is a critical networking model for operators because it translates network and service objectives directly into network actions, including service/capacity provisioning, assuring and scaling. Thus, intent-based networking addresses end-user requirements (the "intent") while abstracting complexity.

To what extent do you agree with each of the following statements?

AI and ML

AI and automation are not synonymous terms, but they share a highly symbiotic relationship. AI helps improve the output of automation. At the same time, automation (and the observation of the outcomes of changes) aids AI in doing its job by providing the data needed for model training and tuning to improve accuracy. Similarly, ML is an important subset of AI that allows computers to learn from and make predictions or decisions based on exposure to datasets.

Heavy Reading research shows that AI and ML are fundamental to operators' transport automation strategies, but — as with intent-based networking — there is still a long way to go before adoption becomes ubiquitous. Of the operators surveyed, 84% are currently using AI/ML in some form. Parsing this number further, 46% of operators are using AI/ML for some network operations and seeing benefits, while an additional 30% are in the experimental phase with AI/ML tools. As expected, very few report extensive use today (just 8% report AI/ML use in most or all processes).

To what degree have you implemented effective AI/ML tools to improve your teams' network operations effectiveness?

Accurate, reliable and timely data inputs are essential for the success of AI/ML, and machine-generated output is only as good as the inputs into the system. Unreliable output can quickly erode the potential benefits of AI/ML as well as destroy operator trust in these emerging technologies and delay further adoption.

Automated active assurance

One data source that can complement traditional network telemetry and monitoring in transport networks is active assurance. It involves proactively testing networks, including running tests on-demand and running synthetic tests to simulate user experience.

The Heavy Reading survey data shows that, while not widely used today, automated active assurance is an important element of operators' transport automation plans. Within one year, more than half (54%) of respondents expect to have automated active assurance. And within three years, 96% of respondents expect to have it — the greatest percentage of the four networking innovations surveyed.

We also note that automated active assurance ties in closely with intent-based networking and root cause analysis and remediation in ensuring end-to-end service quality (both of which are also included in the survey question). Ideally, all three will be used together throughout the lifecycle of a service. The closely aligned timelines identified by survey respondents suggest that they do, in fact, plan to use these elements together.

What timeframe are you planning to deploy the following networking innovations?

Timelines in surveys can be tricky, as operators tend to underestimate the challenges that can slow progress. But they are a highly reliable predictor of where operators want to go in the near to medium term. Collectively, the data from this year's survey indicate that transport networks of the future will become more automated and intent-driven — with extensive use of AI and ML — as technologies continue to mature.

Looking for more information?

This blog is sponsored by Juniper.