HFC Crystal Ball: Methods to Predict Traffic Patterns of Broadband
Wednesday, October 13, 2021: 9:30 AM - 10:30 AM EDT
DescriptionOut of any evolution to the cloud comes huge volumes of useful data that can be used in valuable ways. In the case of broadband network operators, the methodology to predict times when data volume will hit network nodes hard, or, to establish when the “Net Present Value (NPV)” and “Internal Rate of Return (IRR)” of a virtualized environment is right-sized for automation is the precursor to making big data operationally valuable. That’s the gist of this workshop, moderated by Ciena’s Fernando Villarruel. Cox’s Wei Cai opens the discussion with details of a traffic prediction pipeline the company developed for big data, by time, and using cluster-based techniques. Next, Comcast’s Brian Gray shares the journey of a complex cloud workload, from manual to automated and self-managed; importantly, the group that developed the mathematical models deliberately characterized them in well-known financial lingo, like NPV and IRR.