SAN DIEGO -- Cisco has launched a cloud-based AI and machine learning analytics platform that can draw data from any portion of the enterprise -- the data center, campus and WAN -- to provide network and application visibility and insight through a subscription service.
The company introduced its ambitious Cisco AI and machine learning platform during the opening keynote on Monday at the Cisco Live user and partner conference. The analytics service would slice and dice aggregated data from subscribers, as well as metrics from a single network.
"It sounds like Cisco is planning to crunch data from everywhere and in all forms," said Shamus McGillicuddy, an analyst at Enterprise Management Associates. "Its data will be more diverse and broader in scope than many of its competitors."
The result would let customers determine what is typical network and application performance for them. Subscribers of the Cisco AI and machine learning service could also use intelligence from the aggregated customer data to compare their network performance with that of others.
Cisco plans to offer the service first to subscribers of DNA Advantage, the mid-tier license for DNA Center. The latter software is a central network management console that lets customers set policies that govern the operation of a Cisco-powered network fabric.
Cisco AI, machine learning analytics crosses all network domains
The new analytics service, which Cisco plans to make available in the summer, will work across several of its network domains. They include SD-Access, which lets engineers set policies that control employee access to the network; software-defined WAN in the campus; and Application Centric Infrastructure (ACI), a policy-based software-defined networking architecture for the data center and public cloud. ACI includes integration with Cisco's AppDynamics application monitoring software.
A closer look at the SD-WAN piece can provide an example of how the new service would work through its integration with ACI and SD-WAN. The latter could prioritize specific data center application traffic to ensure that it meets an acceptable level of service.
"This will require not only the visibility but also a policy translation capability that ensures that data center policies are properly executed in a WAN context," said Brad Casemore, an analyst at IDC.
The use of aggregated customer data is not new to the industry. Network analytics startups, such as Nyansa, provide a similar service. However, the difference is in the "sheer number of customers," McGillicuddy said.
"As adoption rolls forward, [Cisco] will have much more customer environments to analyze than anyone else -- if they can scale it out," he said.
Sharing data is optional
Cisco customers will have the option of not sharing data with the service. However, refusing to participate in the Cisco AI and machine learning platform would make the intelligence from the service "less knowledgeable," McGillicuddy said.
Large enterprises are usually skittish about having performance data leave their networks for a cloud-based service.
"I've heard this from customers of other vendors who offer cloud-based analytics," McGillicuddy said. "There are some conservative industries that are slow to adopt this sort of thing."
In other Cisco Live news, the company announced the release of the IR1101 industrial edge router that supports Cisco's SD-WAN for managing industrial IoT traffic from remote locations.
Also, Cisco plans to release two ruggedized products for use by manufacturing, oil and gas companies, and utilities. The new devices include the Catalyst IE 3400 switch, available in the summer, and the IW6300 Heavy Duty Series access points, available in the fall.