How to integrate computing and networking services autonomously through AI

Commercial context

The emergence of generative AI (genAI) and large language models (LLMs) has fueled the expansion of applications in domains such as AI generated content and LLM-based training and reasoning. Yet these applications pose new demands on computing and network resources and operations, such as heterogeneous resources, distributed deployment, and non-deterministic computing patterns. Many businesses therefore lack the infrastructure, quality data, model training capabilities, specialized tools and skills to apply them effectively.

CSPs and their partners can help businesses overcome these challenges by providing integrated computing and networking services - but existing siloed resource management and inefficient operations are insufficient to support such new AI businesses at scale. CSPs’ ability to provide sophisticated business customization and cross-professional operations efficiently can be hindered by operational models and IT infrastructures that are predominantly manual and dependent on the expertise of business personnel and specialized maintenance teams.

GenAI and LLMs also open up opportunities for CSPs to significantly enhance operational efficiency and develop innovative business models. Through AI and automation, CSPs can streamline the allocation of ICT resources and deliver seamless, integrated services to democratize genAI and its application for industries. This approach could underpin the delivery of cost-effective and highly customized solutions for computing and networking integration, positioning CSPs for strong revenue growth.

The solution

The Catalyst, which won a 2024 Catalyst Award in the Outstanding Catalyst - Innovative & Futuristic category, builds an AI-native computing force network (CFN) operation system which empowers CSPs to efficiently schedule computing and network resources across cloud-edge. It enables CSPs to provide one-stop intelligent business marketplace through intent-driven CFN services. Employing advanced LLMs, multi-agent collaboration and automated code generation, the solution uses genAI to comprehend, inquire about and supplement the customer's business scenarios and stipulations. Equipped with a thorough understanding of computing and network resource information, the solution then autonomously crafts a bespoke computing and networking integration proposition that precisely meets the client's specific requirements. The client can then refine the proposition using an intuitive chatbot customer interface, which replaces the conventional product catalog.

The solution is designed to plan and execute comprehensive integration tasks throughout the lifecycle of business intents, assign specific tasks to dedicated computing and networking professional agents, and automatically orchestrate procedural business process management strategies to completion. “This Catalyst aims to address the challenges and capitalize on the AI boom by offering a secure, adaptable computing and network service that seamlessly aligns with the customer intent throughout the lifecycle,” explains Jianjian Yang, director, China Unicom. “It integrates diverse resources from cloud providers, telecom clouds, and third-party services. This will allow CSPs to orchestrate a unique ecosystem by a revenue sharing business model.”

As well as drawing on TM Forum's three-tier architecture for autonomous networks, the Catalyst employs TM Forum interfaces for intent and service orchestration, and TM Forum-standardized resource and service management components.

Applications

The solution developed by the Catalyst is designed to enable enterprises to rapidly and affordably use computing and networking services, while bypassing the conventional stages of planning, designing, and deploying resources. Through direct engagement with CSPs, services can be initiated in a matter of minutes. Through the work of the Catalyst, a CSP will be able to offer a comprehensive, one-stop solution underpinned by shared, reusable resources and services. As a result, enterprises could save 30% on hardware and 50% on skilled labor costs.

To demonstrate the value of its approach, the Catalyst has delivered a comprehensive AI interview service to recruitment firms, encompassing initial customer intent assessment, solution formulation and automated cloud-edge deployment for large model training and reasoning, as well as ongoing business support and self-service functionalities for the clients.

Wider value

The Catalyst is paving the way for CSPs to achieve advanced levels of autonomous and orchestrated operations in computing and networking services. By enhancing CSPs’ operational capabilities to autonomous operational level 3.6, the Catalyst solution promises to automate 60% of processes, improve resource utilization by 40% and cut operational costs by 50%.

More broadly, the Catalyst hopes to revitalize the entire computing resource pool for the digital ecosystem and society. It is aiming to stimulate the growth of intelligent business and enable the flexible allocation of computing and networking resources, thereby enhancing resource utilization, and reducing energy consumption to support sustainability. At the same time, the solution can help to further integrate the computing and networking industries by establishing innovative cooperative models, offering new growth opportunities for partners, and driving the development of intelligent business practices.

To learn more about how this Catalyst is using AI and automation to streamline the allocation of ICT resources, please see the project space on the TM Forum website here.