Proactive Service Assurance via Closed Loop Predictive AIML

Business Drivers For the Catalyst Complex networks require automation Operational Efficiencies • SDN/NFV and

The Conceptual Solution © 2019 TM Forum | 3

Conceptual Solution Proactive Service Assurance via Closed Loop Predictive AI/ML Predict network events that

Concept of Closed-loop Automation via AI Principles • • • Champions Du: Service Provider

Potential Business Benefits How many staff do you have in contact centres? 10, 000?

Completing the Catalyst Project © 2019 TM Forum | 7

AI Service Assurance Solution Architecture © 2019 TM Forum | 8

General Observations • • 80% of alarms were made up of approx. 15 types

Predictions • • More accurate for certain types of problem – Achieved 99. 98%

Automation Scenario Examples • • Pro-active Traffic Management Actions to take if accuracy of

Thank you! Contact us to learn more: • Aravind Chennuru, Enghouse Networks Aravind. Chennuru@enghouse.

Slides: 13

Download presentation

Business Drivers For the Catalyst Complex networks require automation Operational Efficiencies • SDN/NFV and 5 G will make it more difficult to effectively manage the network infrastructure manually • Complex eco-systems with more devices, changing low latency, high bandwidth services and a changing network with shifting services • ARPU capping in most regions, additional investments necessary in the network • Continual data growth of 50 -60% per annum Shift in Customer Expectations • High demand for network availability and service quality • Network speed, Time to Market, Time to Resolve etc. become the new differentiators © 2019 TM Forum | 2

Conceptual Solution Proactive Service Assurance via Closed Loop Predictive AI/ML Predict network events that are highly likely to cause service degradation or interruption, and automatically respond with pre-emptive actions to reduce or prevent occurrences. © 2019 TM Forum | 4

Concept of Closed-loop Automation via AI Principles • • • Champions Du: Service Provider use-cases and network data Sri Lanka Telecom: Service Provider usecases and network data Our Catalyst project demonstrates advancement toward fully automated Service Assurance for evolving networks in digital transformation • The project works with actual network data and utilizes AI, predictive analysis, and knowledge-based self-learning ____ The output delivers automated issue resolution through machine-to-machine or machine-to-human interaction • Participants • • • Enghouse Networks: AI powered Service Assurance Net. YCE: Automated Provisioning in response to network events and AI/ML output Florida Institute of Technology: Data science and analysis © 2019 TM Forum | 5

Potential Business Benefits How many staff do you have in contact centres? 10, 000? Reduced Service Degradations & Interruptions • The predictive element helps avoid & reduce degradations & interruptions Improved End-User Experience • Minimized frequency of customer contacts reduces churn • Reducing Mean Time To Repair (MTTR) • Network queries: 3 -10% of all CSP customer contacts What is your average truck roll cost? $500? How high is your network availability? 93%? Accurate problem diagnosis • Reduces Mean Cost to Repair (MCTR) • Reduces unnecessary truck rolls Improved Network Availability • Reduces mean time to repair (MTTR) © 2019 TM Forum | 6

Predictions • • More accurate for certain types of problem – Achieved 99. 98% accuracy on threshold alarms by time and severity Volume of alarms has a huge impact of the predictability of the fault – 74% in the worse case scenario © 2019 TM Forum | 11

Thank you! Contact us to learn more: • Aravind Chennuru, Enghouse Networks Aravind. Chennuru@enghouse. com • Wim Gerrits, Net. Yce Wim. Gerrits@netyce. com • Phil Chan, Florida Institute of Technology Pkc@cs. fit. edu © 2019 TM Forum | 13