Actionable versus Predictive Intelligence: Nuanced Approaches to Operational Analytics

Actionable versus Predictive Intelligence: Nuanced Approaches to Operational Analytics

  • April 2015 •
  • 24 pages •
  • Report ID: 2875518 •
  • Format: PDF
This report will focus on the NetBoss solutions for actionable intelligence; Nokia’s predictive operations solutions; IBM’s and Nokia’s flirtation with cognitive networking; and Ericsson’s experience-driven platform approach to expert analytics.

The stakes have never been higher for communication services providers (CSPs). Struggling against physics, economics and well-funded, lightly-regulated competitors, CSPs continue to build networks and operate businesses that must support the insatiable demand of billions of technology users, and the continual introduction of innovative applications. The higher the stakes, the more intelligence CSPs need to make informed decisions about how to manage their networks and businesses. And, the more complex their operations become, the more automation CSPs need to incorporate. Automation, in turn, requires yet more intelligence.

So, the big question at hand is: From where will all of this intelligence come? Surely, CSPs do not lack data; but turning their data into useful intelligence takes increasingly sophisticated analytics. But, there are other questions as well:

- How does the intelligence differ for decisions that must be made on the spot versus decisions affecting more long-term planning?
- Who decides which data is useful for developing the intelligence?
- What actions can be confidently automated?
- How can a CSP foresee all the consequences of an automated action?
- Can a CSP trust its operations, its network and its business to a set of algorithms and a machine?

To this last question, some network management suppliers of assurance and analytics solutions would have CSPs think they can put operation of the network into the “hands” of autonomous, selflearning machines. Stratecast believes such overconfidence is misplaced and unproductive; a position that will be borne out in this report. However, these same suppliers and others are making notable progress in bringing predictive analytics to the forefront of the network operations environment.

They are doing it with machine-learned command and control capabilities that, in many ways, demonstrate superiority over what the human element can accomplish for a growing number of selective network management needs. Machine-learned control also brings to light new business management opportunities not thought possible before now.