Predict Real World Outage using Splunk MLTK [Splunk Machine Learning Toolkit]

Do you want to predict an outage before it happens? Are you wondering how to pursue the incremental journey to Artificial Intelligence Operations (AIOps)? This case study will reveal a real-world use case from T-Mobile USA and show you how to predict cell tower congestion in advance using Splunk Machine Learning Toolkit. In the age of binge watching on cell phones and wireless broadband services, cell congestion reduces speed and reliability and results in buffered video streaming and/or dropped calls that dents the use of services and the revenue. Building forecasting models for congestion requires correlation of several parameters including seasonal variations. Doing this on a large scale in real time takes significant resources. In this session, attendees will learn about the journey to build this predictive capability, including data analysis techniques, machine learning algorithms, benefits, and lessons learned.

Speaker(s)

Vijay Veggalam, Member of Technical Staff, T-Mobile

Gintaras Gaigalas, Sr. RF Engineer, T-Mobile

Slides PDF link - https://conf.splunk.com/files/2019/slides/IT1722.pdf?podcast=1577146258