Several of TM Forum’s largest telco members worldwide have kicked off collaboration on a new industry-specific data reference architecture that aims to encompass both new AI-enabled business models and the running of networks.
In total 18 communications service providers (CSPs) were present at the launch of TM Forum’s Modern Data Architecture for Telecom Operations Project in early December, including Axiata, China Mobile, MTN, Orange, Telefonica, Telenor, Vodafone and Zain. They were joined by a lineup of suppliers and systems integrators that included Cloudera, Huawei, Accenture, Ericsson, and Infosys.
The new project reflects network operators’ need to build AI-driven autonomous processes, to put data and AI at the core of operations, and to bring external data into their meta-ecosystem as their ecosystems expand. It also tackles the question of what a modern data architecture should look like for AI-enabled telecom operations.
One thing was clear from the first meeting: many telcos see a very real and urgent need to modernize their data architecture and they want to play a key role in shaping it. This marks something of a shift, as until now vendors have done much to model existing telco data architecture. At the same time, the insight and experience of suppliers and SIs within the group brought tremendous value to the conversations about the direction of travel. Indeed, if the ideation process is anything to go by, then this initiative will be a game changer for operators, solution providers, and integrators alike.
Some key takeaways:
- SIs and Platform Providers recognize that data and AI is increasingly central to telecom industry operations and that platforms need to be accessible, AI-ready, and well governed, supporting all network types and storage on distributed public and private cloud infrastructure, using lake, warehouse and lakehouse alternatives, and allowing exposure through AI and development functions to enable automation.
- CSPs see an urgent need for a telco specific data reference architecture that looks at data beyond SID compliance. Telcos are keen to see architectural references run not only on telecom networks, but to effectively allow them to manage new business models.
Key questions included:
- What challenges should a modern data platform address?
- Can the modern data platform become a first-order environment, where network data is published directly, rather than a second-order platform primarily dependent on data from other applications like service assurance, partner and Customer Experience Management and billing?
The focus of this new TM Forum project is, therefore, to collaboratively explore and clearly articulate:
- What the modernized data architecture should look like for the telecommunications industry, and
- Why does the modernized data architecture matter (i.e. the impact / extent to which the modernization will solve key issues for the telecommunications industry).
During the project kick-off CSPs raised nine challenges, which project members have prioritized as follows, based on their business impact:
- Solving the technical debt challenge: Modernizing the data architecture must ensure telcos can eliminate technical debt in telecom operations
- Ensuring business-oriented alignment of the data architecture: There is a need to move away from data warehousing / centralized data when modernizing telecom data architecture because data lakes are not scalable especially when running a group of telco entities or a meta-ecosystem. This in turn raises the question of what the decentralization of data means when transforming from a legacy architecture.
- Modernization of data architecture for telecoms MUST respond to the direction in which telcos are taking their digital transformations.
- Harmonization of telco (OSS/BSS) data and non-telco data with external data
- Integration of Open API use-cases which impact customer experience
- Enabling GenAI. This entails addressing what architectural adaptation and change GenAI requires and how to enable the use of unstructured data in Autonomous Networks
- Enabling standardized data products and access to data sets to create good AI use-cases for the customer, which involves addressing what types capability are needed to build data products.
- Model Federation and developing the capacity to manage multi-model and multi-party data ecosystems.
- Articulating the role of the knowledge and intelligence management platforms in the end-to-end ecosystem autonomization of operations. Questions here include how should applications like real-time AI-based anomaly detection, service assurance, partner and customer management and other AI-enabled use-cases of data be interacting within the Architectural framework of such a first-order data platform? And where are the points of convergence within the context of ODA?
TM Forum Members can view or join the project here.
The Modern Data Architecture Project will be a big focus at Accelerate 2024 in Cascais, Portugal, 5th – 8th February 2024, for which registration is now open.
