logo_header
  • Topics
  • Research & Analysis
  • Features & Opinion
  • Webinars & Podcasts
  • Videos
  • Event videos

Using data, AI and automation to control services and the network

It is difficult for telcos to integrate the network and business areas; this proof of concept Catalyst project seeks to merge those processes and automate them.

Annie Turner
09 Sep 2020
Using data, AI and automation to control services and the network

Using data, AI and automation to control services and the network

Network as a service (NaaS) is the key to re-establishing operators as the owners of the relationship with B2B customers. To succeed in this, they need closed loop automation and AI enabled operations (AIOps) to automatically meet and maintain customers’ service needs dynamically. This means automation must span the entire hybrid network, encompassing physical and virtual elements. It also means integrating the operational and business support systems (OSS/BSS) – a goal long strived for by operators.

In this Catalyst project, Boosting AIOps for full hybrid NaaS, the champions are American Tower Corporation (ATC) and Telefónica, supported by participants Everis, Verbio, and Vlocity, a Salesforce company. In this instance, ATC is the network provider and Telefónica is the retail service provider to enterprise customers.

Dahyr José Vergara Suárez, Solutions Manager for Telecoms at Everis and a Catalyst team Co-leader, explains, “It is difficult to integrate the network and business areas but in our Catalyst we are merging those processes and automating them.”

Importantly, the team wanted to ensure what it developed could be transposed to other services, whether wavelength as demonstrated in the Catalyst, mobile, FTTx, etc. In preparation, the team analyzed many case studies and became familiar with TM Forum’s TM Forum Business Process Framework (eTOM) (also known as eTOM) processes. It also used the Open Digital Architecture to help with the BSS and OSS integration.

Service performance


The central issue is that while there are many models to predict events such as an overflow of traffic, and although equipment vendors can balance traffic at the device level, operators themselves have never been able to do the same at the service level in a smart, cost-effective way. This can have a profound effect on customer experience.
Vergara Suárez stresses, “Data is the key component and data management must be inside the architecture for operations to be data driven and implement AIOps. You cannot do anything without data, a single data model, and an architecture that enables you to integrate all the information from different sources.”

Previously, and separate from the Catalyst, Telefónica and Everis developed Fast OSS. It ingests, normalizes and models data from multiple sources, such as the underlying cloud architecture, the servers, memory and network services, the virtual network functions and software-defined network controllers.

In the Catalyst, the Fast OSS is complemented by UNICA Next which analyzes the data and acts on it. For example, if the Fast OSS needs to scale up a function in the evolved pack core when a serious overflow of traffic is predicted, capacity can be duplicated or halved or whatever is necessary to maintain the right level of service – dynamically and without human intervention.

This can be applied to any network function using the UNICA Next architecture, taking it up to the service level – and even to the customer level ­– because everything below in the network is automated. The provider could even smartly restart a site as an operator currently does from a network operation center.

A single data model


Having all the data within one data model means that if a server fails, it is easier to determine what the impact is and address it dynamically and automatically. The Fast OSS and UNICA Next working together can create tickets automatically, correlate information from different tickets with different network events, create new events from previous events, or create an event from network performance metrics.

TM Forum’s Open APIs are the key to this integration. Vergara Suárez says, “We found that using the APIs is really good experience. Normally we have a to modify them a lot because of aspects specific to customers, but in the Catalyst, we did the opposite: We found that by doing little tweaks to the platforms and the systems, we can comply with the standard [API], and be a lot more interoperable, easily”. Everis intends to recommend this approach to customers in future.
Vergara Suárez notes, “This Catalyst meets the champions’ needs and every participant is key to making that happen.”

Blockchain


This unusual approach to integration helped to design the resource pool to work with blockchain from scratch, using the APIs as the foundation. The blockchain establishes communications between the provider and the retailer. Vergara Suárez explains, “Normally there is a team set up on both sides of the transaction, because you need an account manager and a team as one person cannot manage everything related to billing and service level agreements.
“We automated the whole contract between the two without human intervention. Whatever is recorded in the blockchain stays in the blockchain; it is communicated to different solutions and cannot be modified.”

Some team members want to take various strands of the work in this Catalyst further. It is not yet clear will mean another phase of this Catalyst or continuing parts of the work in other Catalysts.

Click here to watch the video of Boosting AIOps for full hybrid NaaS demonstration, as part of the Catalyst Digital Showcase, and watch a general overview of the project in the video below.