Introducing predictive analytics and deep learning, systematic control of software changes, and speedy network-level fault isolation and recovery

Photo Credit : AP,

To help operators cope with the demands inherent in connecting an increasingly complex world, Ericsson is introducing a new approach to Support Services.

It involves complementing Ericsson’s existing Support Services with predictive analytics and deep learning, introducing more systematic control of software changes, and speeding up network-level fault isolation and recovery. Ericsson will also provide actionable intelligence that enables operators to achieve new levels of network stability.

As businesses of all kinds reinvent themselves to realize the full potential of data, operator networks are coming under increasing pressure to connect anything and everything. The introduction of different use cases such as video and audio on demand, self-driving vehicles and factory automation will require network latencies under 10 milliseconds, with no room for error.

Fredrik Jejdling, head of Business Area Networks at Ericsson, says: “As operators face increasing network complexity with the introduction of new use cases, they must become more proactive. Through close collaboration between our support engineers and our customers’ operations – and by making use of automation, machine learning, and other artificial intelligence techniques – we’re putting the zero-defect network vision within our customers’ reach."

Rodrigo Orozco, head of Network Operations at Entel Chile, says: “We are working with Ericsson to explore the power of data analytics to enhance network operations. Together, we were able to prevent approximately 85 percent of critical incidents in the network. This proactive approach significantly reduced the impact of any network issues on our end users.”

In other trials and early deployments, Ericsson has enabled operators to achieve a near-perfect success rate in handling software changes and drastically reduced emergency recovery time from four hours to just 60 minutes. In one example, a European operator was able to reduce network incidents by 30 percent despite tripling the total number of upgrades per annum.