Closing the language and operations gap between IT and network engineers

There is a communications gap between communications service providers’ (CSPs’) IT and network engineers which costs the companies a great deal of time, resources, money and effort. This ambitious and innovative proof of concept Catalyst project builds on a previous phase to develop speech technology that can automate operations and business support systems (OSS/BSS) functions and processes to help close that gap – and save operators time, money, resources, and effort.

Sri Lanka Telecom is championing the Catalyst, SINA for field OSS – (SINA) Speech interpretation network assistant – Phase II, supported by participants Clarity, Enghouse Networks and the Florida Institute of Technology.

Aravind Chennuru, Software Architect, Enghouse Networks and Catalyst project lead, explains what was achieved in the first phase, titled Proactive service assurance via closed loop predictive AI-ML, in the video below.

In short, the team applied machine learning to trouble ticket resolution using a knowledge repository to speed up the fix. It also looked at how predicting network capacity needs can enable CSPs to auto-provision bandwidth. They were able to automate – that is, close the loop in the machine-learning system – by automatically correcting problems based on the accuracy of the predictions.

Avoiding complexity

Having proved that data can be used with AI to successfully assure service in these ways, the team wanted to lower the barriers to implementing AI by abstracting service assurance applications. Chennuru says, “We introduced something called SINA for Speech interpretation network assistant because we wanted to enable field and network engineers to talk to a system without knowing the complexities involved. It makes their jobs easier and they can just get on with their tasks”.

Today field engineers have little understanding of operational systems and typically call an operational expert when they require help. Chennuru comments, “There's too much time wasted, and too many people and resources to do the same job.”

He stresses SINA is not just about making information accessible to network or field engineers, but automating tasks and entire process work flows. The team uses the example of a field engineer arriving at a site and preparing to start work. They need to notify software operations and the service operations center so that the monitoring of alarms is suspended while the work takes place.

“In our Catalyst, all this is automatically taken care of because you create a process and the process activates itself on the user’s command. So the time to perform complex operations will be reduced drastically,” according to Chennuru.

Next steps

The team used the TM Forum Open Digital Architecture (ODA) as a reference but wants to augment it by segregating data and functions, such as fault management, performance management, the ticketing system and configuration management, thereby making the systems independent of each other.

Chennuru explains, “Today we have a direct communication between these systems using north bound interfaces, but that doesn't work when a system is out of support or at end-of-life. This is a huge cost to CSPs. Our proposal makes the systems independent so we can ‘plug and play’ them without affecting the entire network; such integration in future will cost the CSPs less.”

The team used the Forum’s Open APIs (specifically TMF621, TMF640, TMF642, and TMF649), to which it wants to add features to support Apache Lucene-based queries. Lucene is a simplified query language. Chennuru states, “We were hoping we could use Open APIs out of the box but we had to modify them to make SINA work”.

The team intends to give SINA more features; at the moment, it integrates the fault, performance and trouble ticket management system.

Next phase

The intention is to create a framework or toolbox that includes all the proposed changes to the Open APIs and ODA changes, plus the rules necessary to build the SINA application for Phase III which it expects to demonstrate in summer 2021. Chennuru says, “We intend to build an actual product as an example of what you can do using the toolbox”.

To this end, it will demonstrate a more complex version of SINA that integrates a fiber management system, which has five or six operational heterogeneous systems, so the team will be working on many more integrations.

Chennuru concludes, “The idea is that people will be able to use elements of the toolbox to build an app for CSPs so that they can easily use AI for predictive provisioning and to activate automated operations by text or voice command”.

Proactive Service Assurance via Closed Loop Predictive AI/ML was shown at Digital Transformation World in Nice in 2019, watch the brief video here.

This demonstration video shows SINA for field OSS – (SINA) Speech interpretation network assistant – Phase II in action as part of the Catalyst Digital Showcase.