This is an excerpt from our Trend Analysis Report on AI. We surveyed 187 executives from 76 communications service providers (CSPs) operating in 51 countries, and 115 executives from supplier companies. We also conducted dozens of interviews with respondents.
Telecom networks are complex, and the transition to cloud, virtualized functions and software-defined networking is increasing the complexity. This makes it difficult to develop products for specific network use cases. “The level of maturity is not there,” says Utpal Mangla, a partner in IBM’s Telecommunications, Media and Entertainment Industry division.
Complexity needs AI
But it is precisely the complexity that makes AI so promising. As the internet of things (IoT) grows, network and service management must be zero-touch, because it simply isn’t feasible for manual processes to support the volume and velocity of changes that must happen. AI and machine learning will be required, which is why it is paramount that CSPs embrace the technologies.
“A network servicing 10 million endpoints and 10,000 nodes could see these numbers increase by up to five times by 2020. In terms of incidents per hour, this would lead to a 25-times increase from 400 incidents to as many as 10,000 per hour – an amount that is impossible to handle manually.”
World Economic Forum’s Digital Transformation Initiative Telecommunications Industry white paper
Autonomous networks have potential
Early use cases
CSPs can buy some network management AI products and services today. For example, IBM Watson Field Service Advisor helps technicians resolve field service requests by making specific recommendations about how to resolve problems. Accenture calls this type of application “AI-powered over-the-shoulder support”.
From NOCs to SOCs
Other products help CSPs transition their network operation centers (NOCs) into service operation centers (SOCs). While a NOC is where administrators supervise, monitor and maintain a telecommunications network, a SOC uses analytics and AI to deliver closed-loop automation. This enables CSPs to quickly detect, diagnose and recover from service-impacting issues without human intervention.
Telefónica has launched pilot SOCs in Argentina, Chile and Germany with plans to transition to SOCs in all its core markets. Enabled by AI, the SOC capabilities will empower the company to capture in real time the true quality of customer service experience.
Telefónica says the application of AI to networks will maximize capacity and solve any problems before end users even notice anything. Through the use and interpretation of data in the network, the company aims to transition from scheduled network maintenance to predictive and proactive maintenance.
It takes AI to TANGO
South Korea’s SK Telecom has launched what it calls Telecom Advanced Next Generation Operational Supporting System (TANGO) to improve both traffic management and network operations more broadly.
The company spent two years developing TANGO which uses big data analytics and machine learning. It breaks down network traffic information by area and period, detects abnormalities in the network, and provides the best solution to fix problems.
Vodafone, meanwhile, already has bots that handle network alarms and is starting to apply automation and machine learning to help prioritize trouble tickets.
Dabbling in drones
A more experimental, and ambitious, application for AI involves the use of drones to inspect base stations. Many large mobile operators partner with contractors for drone inspection, but they want to improve the process using AI.
A team at AT&T Labs, for example, is working on a deep learning algorithm that can analyze video footage looking for defects and anomalies in base station infrastructure. The goal is to use fully automated drones to inspect and repair its 65,000 cell towers.
Proactive capacity planning
With data traffic growing by 50 percent a year in most countries, CSPs must continuously add capacity to their networks. AI-powered tools are under development to forecast growth in capacity and predict failures in the network.
AI may also have a role to play in analyzing traffic flows as voice calls are handed from one operator to another, and in financial settlements between operators.
Indeed, PCCW is already testing a system with one of its partners to use heuristics – the calculation of trade-offs between optimality, completeness, accuracy and time – to determine whether the discrepancy in traffic flows between partners is large enough to warrant a commercial settlement.