By deploying analytics and business policies, automation can make network change decisions as if the customer had made those decisions themselves.
DENVER -- NFV & Carrier SDN -- Using analytics, Level 3 can provide network services that deliver quality as good as if the enterprise customer had been managing the network personally.
"I think that's probably the most exciting arena," David Siegel, vice president for data services product management at Level 3 Communications Inc. (NYSE: LVLT), said during a Q&A presentation. Automation can eliminate the need to wait for the customer to review analytics reports, and call Level 3 to order a change. Instead, the system acts automatically.
"The system can make the decision the customer would have made," Siegel said. Later, the customer can review the decision and, if it's not to the customer's taste, the customer can review and change the business policies that led to the decision, Siegel said.
Figure 1: Analyze This Level 3's David Siegel talks analytics with Light Reading's Carol Wilson.
Level 3 started using analytics six to seven years ago, with Hadoop clusters providing big data analytics to process netflow data for network planning, providing cost models for commercial analytics to assess profitability, Siegel said. (See Level 3: Analytics Unlocks SDN Potential.)
Level 3 processes quoting data to reoptimize pricing to achieve business goals -- improving topline revenue, margin, or growth, Siegel said. The network operator also offers analytics products to customers. A VPN customer can get a hardware appliance to use analytics to improve network performance, he said.
Security is another "exciting" use case for analytics. The more data that goes in, the better the results get, with fewer false positives and fewer missed attacks. Level 3 integrates DNS data from millions of users and data from its content delivery network as well. All of it is integrated and bundled together into the Level 3 Active Threat Intelligence Service, Siegel said. (See Level 3 Makes Security a Network Matter.)
Level 3 deploys analytics as part of its SD-WAN offering. Analytics capability is bundled into vendor-provided engines, providing capabilities similar to Adaptive Network Control, which is Level 3's SDN offering, launched in 2015. Customers can control their networks using business policy. For example, customers can say they want to update bandwidth by a certain amount, at a particular location and time, for a specified duration. Or the policy can be event-based -- for example, when bandwidth utilization gets above a certain level, the circuit increases, paid for by the hour until usage subsides. The data governing those decisions is very simple, married with business policy to create value for the customer, Siegel said. (See CenturyLink Sets Sights on SD-WAN and Level 3's SDN Play Focused on Hybrid Clouds.)
But analytics is not without problems. The sophistication of packages means that out of hundreds of customers, literally only two or three are skilled enough to use analytics, Siegel said.
"The range of troubleshooting possibilities is endless, but you have to be a network engineer in order to use the information, and the interface is so full of information you have to be trained on it," he said. "A lot of people wouldn't necessarily find the value of doing that."
It's too much trouble for many of Level 3's enterprise customers. "They may not have the people or resources. They might want us to do it on their behalf," Siegel said.
Find out more about analytics on Light Reading's Analytics Channel.
Analytics needs interfaces suitable to all levels of sophistication. Customers want to specify outcomes, and not have to diagnose what happens on the network.
Also, talent is a problem for the whole industry, and analytics sometimes requires going outside the engineering pool to find expertise. For example, when analyzing financial data, or broad, general company data, people in the finance organization know how to use it.
Level 3 has also hired microeconomists who know how to use system analyst tools. Microeconomics is "a branch of economics that studies the behavior of individuals and firms in making decisions regarding the allocation of limited resources," according to Wikipedia.
Related posts:
— Mitch Wagner, , Editor, Light Reading Enterprise Cloud
About the Author(s)
You May Also Like