Nokia today formally unwrapped a series of artificial intelligence (AI) and machine learning features that are already being used by operators in the Asia-Pacific region, Europe, and North America. Nokia AVA 5G Cognitive Operations is designed to increase automation of network operations by maintaining service requirements and predicting challenges before they occur.

The tools can also help operators make better decisions about network slices and meet service requirements, according to the Finnish vendor. The software is multi-vendor by nature, according to Dennis Lorenzin, head of network cognitive services at Nokia, and therefore not exclusive to operators that use Nokia gear.

“The functionalities are working as well for earlier technologies than 5G, though they were developed having 5G expected complexity in mind,” he said. By combining data science, machine learning, and AI, the service can anticipate network failures up to seven days in advance and solve those problems up to 50% faster, according to Nokia.

“These capabilities are already being used by Nokia and operators to manage several networks around the world,” including an operator in Japan that is using AI to fix non-performing cells in minutes, Lorenzin said. “With the old methods this took several hours and only yielded an 80% success rate.”

Nokia AVA 5G Cognitive Operations is built on Microsoft Azure, but it can be extended to other cloud providers to meet specific needs of operators, he explained. 

“We wanted to address specifically the requirements of those customers who are willing to adopt webscale capabilities to support their future network evolution and operations,” Lorenzin said. He described the effort as a combination of Microsoft’s infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS) capabilities, and Nokia’s expertise in telecommunications-specific applications.

“Besides the IaaS/PaaS services provided by Azure, or other webscale [providers], we have developed the solution in house adopting open source components” for features like data communication and visualization,” he said. Nokia has also embedded some third-party services in its software for specific use cases or algorithms. 

“We use a variety of machine learning algorithms developed by Nokia Bell Labs, as well as those developed by third parties,” Lorenzin said. The vendor is blending a mix of PaaS with open source technology to more easily port the software to other cloud platforms, he explained.

After network operators select the features or use cases they want to deploy, Nokia will work with the operator and its cloud provider to implement a secure data connection between the network and cloud, he said. Machine learning algorithms are then seeded in gradually to be tested against that service provider’s customer data and the business outcomes it wants to achieve.

“Normally the use cases are used to improve the efficiency of operations, reducing human intervention for labor intensive activities, yet not always resulting in a full closed-loop automation, but rather in a streamlined, simplified, and more efficient process,” Lorenzin said. 

He declined to share how much Nokia is charging for the service, but said it will vary depending on the amount of use cases and total size of operations.

AI and machine learning have gained prominence in the industry of late and radio access network (RAN) vendors are following that trend. Earlier this year, Ericsson unveiled a series of updates to its 5G software platform, including an array of network services that make use of AI to reduce opex and glean more actionable insights from the network in real time. 

Nokia joins the AI fray during a particularly challenging time as it struggles amid a heightened period of 5G deployment activity. Much of the company’s current 5G product portfolio carries high costs that have impacted its profit margin, but Nokia’s executive team insists its new equipment that features a new system-on-a-chip (SoC) is expected to reverse that trend with greater performance at lower costs. However, that new hardware won’t reach a majority of its 5G shipments until sometime next year.

Amid that backdrop, the company earlier this month announced that CEO Rajeev Suri will be leaving his position on Aug. 31 and Chairman Risto Siilasmaa will depart the company this month. Suri will be replaced by Pekka Lundmark, who currently serves as CEO of Fortum, an energy company also headquartered in Espoo, Finland.