Cisco: Artificial Intelligence, Machine Learning Key To Networking's Next Evolution

‘Machine learning is going to help us augment our mental capabilities so we can process information faster when there's information overload,’ says Prashanth Shenoy, Cisco's vice president of marketing for enterprise networks, IoT and developer platform.

Cisco Systems is making its intent-based networking portfolio simpler, smarter and more secure, the tech giant said. That means bringing in more artificial intelligence and machine learning and stitching together disparate parts of the network.

At Cisco Live 2019, the San Jose, Calif.-based company revealed the next step in its networking evolution. The original premise behind intent-based networking involved focusing the network on what the business wants to achieve. Cisco has been achieving this through network automation and boosting security across the entire network. But now, the network has grown even more complex and nearly impossible to manage.

"Network management has exceeded human capacity. Now, IT is getting so many alerts that they aren't able to keep up at the pace that the business wants them to," said Prashanth Shenoy, vice president of marketing for enterprise networks, IoT and developer platform for Cisco.

Networking's next stage requires machine learning, Shenoy said. "Machine learning is going to help us augment our mental capabilities so we can process information faster when there's information overload," he said.

[Related: Cisco's AppDynamics Launches Integration Partner Program To Deepen Visibility For IT Teams]

Cisco has been working hard to harness machine learning to help drastically simplify network operations and provide more intelligent and insightful operations, Shenoy said. To that end, Cisco is using AI and machine learning to offer more granular, customized visibility.

"Every network is a snowflake, so yours is going to be very different from a very similar customer. We want to provide deep visibility that is personalized for customer's environment," he said.

Cisco also is using machine learning to correlate the diverse set of data coming from its global networks and large customer base against a customer's individualized network norms to uncover the issues that will have the greatest impact on the network.

"We're constantly learning, so we put that intelligence back into the network so we can be even more predictive in the future," Shenoy said.

After giving users better visibility and increased insight, Cisco is using machine-learning algorithms to perform logical troubleshooting steps that a human engineer would execute to resolve the issue. These steps will help accelerate issue resolution for partners and end customers, Shenoy said.

The latest updates are completely software-based. Customers that want to take advantage of the new updates, or partners managing networks on behalf of customers, won't need to upgrade any hardware but must enable telemetrics on their existing products.

For partners, Cisco has made its platform open and extensible. These features can be integrated into other IT processes or applications, Shenoy said.

"This provides partners more value-added software practices they can build on top of our network stack," he said.

Cisco is also leaning on machine learning to connect once-separate parts of the network, such as WANs or cloud environments, so they can seamlessly and consistently operate for end customers.

"Bringing together these networks creates consistency in the way [partners and customers] can provide security end to end, from the users to the applications," he said.