Monday, May 27, 2024

Communications Service Providers(CSPs) Adopts AI/ML models

Communications Service Providers (CSPs)

CSPs telecom

AI and machine learning (ML) have been progressively incorporated into networks and commercial operations by the telecom sector. To enhance spectrum efficiency, optimise network performance, and improve energy efficiency, for instance, the Open RAN sector has been working on a number of use cases. In a similar vein, chatbots have been developed by telecom business operations to enhance customer engagement, perform sentiment analysis, and forecast churn.

Utilising AI in the Network

AI is already being utilised today to help networks become more automated and efficient. The industry is set to transition from AI-assisted systems, in which AI is utilised solely in a supporting role, to AI-native systems, in which AI/ML models are used throughout the system’s core development and operation.

AI/ML will penetrate every layer of the network stack, offering communications service providers (CSPs) new chances to build highly automated, differentiated networks that perform better, use less energy, and more. AI/ML has the potential to revolutionise every tier of the network stack, including neural schedulers, neural optimizers, and neural receivers, resulting in improved quality and performance.

Additionally, automation is about to take over the whole network lifecycle, from network design to deployment to optimisation. The use cases span from using generative AI (GenAI) to solve network issues using natural language processing (NLP) superpowers to traditional AI/ML network optimisation. Moreover, Communications Service Providers will be able to conduct thorough simulations in the digital realm prior to implementing changes in the actual physical world thanks to network digital twins.

This is only a cursory examination. Communications Service Providers will be able to create automated and distinctive networks thanks to AI.

AI Assisting in Business

AI will be essential to streamlining, updating, and automating CSPs’ business processes. AI has the potential to help Communications Service Providers increase their top and bottom lines, monetize their networks more successfully, and obtain productivity levels never seen before.

Many CSPs have already started their journey with chatbots and call centre copilots, which are enabling new methods of customer service delivery, sentiment analysis, and proactive churn mitigation. To lead the industry in this regard, they are working with Their partners on the TM Forum Catalyst project “AI chat agent: The game changer for telecoms.” Large language models (LLM) trained on telecom data are being used by this Catalyst initiative to enable chatbots that offer prompt, accurate, and individualised customer support.

Improving client experience and generating revenue growth are the two main objectives. By showcasing how LLM-based chatbots can comprehend user intent and set off automatic actions to present customised data plans or start new service subscriptions, The Catalyst showcases the potential of GenAI.

From field support copilots to code development copilots, copilots will increase productivity, and telecom teams will be able to complete tasks much more quickly thanks to new digital assistants powered by GenAI.

In a similar vein, GenAI will allow network operations teams to find anomalies and cut down on troubleshooting time from days to hours or minutes by enabling them to quickly analyse massive logs and sort through a vast amount of signalling message flows.

AI will be essential to network energy efficiency, as practically all Communications Service Providers prioritise sustainability and reducing energy costs, particularly at the edge in (typically) unmanned places. Energy management at edge locations can be automated by organisations by leveraging AI/ML to recognise what equipment is in use and when.

For networks integrating GenAI, significant advances in productivity and efficiency as well as the potential to reduce operational costs will become possible.

Constructing an Infrastructure Ready for AI

In order to fully leverage AI’s business potential, CSPs must prepare their network infrastructure to handle a completely new category of workloads: workloads related to AI.

AI-ready infrastructure will help Communications Service Providers offer new services and establish new income streams in addition to helping them build distinctive networks and increase productivity. CSPs are in a unique position to provide edge inferencing services.

Networks are ubiquitous and the edge of the real world. The need for inferencing will increase significantly as AI spreads throughout various industries. Consider data transmission costs, latency, the ability to reject low value data at the edge, and other factors when evaluating the technical and financial benefits of inferencing closer to the data generation location. Communications Service Providers can provide inferencing services for this new class of AI workloads at the edge with infrastructure that is suited for AI.

Additionally, CSPs are developing AI-factories in numerous industries to provide services for training and fine-tuning sovereign AI models.

In order to construct AI-ready infrastructure, CSPs will need to address a number of issues. In order to properly size the AI infrastructure for training and inferencing, real-world power, cooling, form factor, and other constraints must be met while also accounting for the sizes and complexity of AI models.

With a broad spectrum of Power Edge servers and GPUs, Dell Technologies can handle real-world limits and the increasing performance demands of AI workloads. As an illustration:

Large AI model training and tuning is a good fit for the Power Edge XE9680 with 8x Nvidia H100 GPUs, which can be hosted in national and/or regional data centres (NDC/RDC) owned by CSPs.

  • The PowerEdge R760XA is ideally suited for core network deployments in telecom and can accommodate up to 4x double-wide GPUs or up to 12x single-wide GPUs.
  • When combined with GPUs, PowerEdge XR8000 CPUs can handle medium inferencing workloads and are AI ready to handle tiny edge AI inferencing workloads.
  • These are but a few instances of how portfolio Communications Service Providers may appropriately scale their infrastructure with Google cloud all-inclusive PowerEdge server and GPU.

The technological world is interesting right now. AI is a revolutionary force that will completely change the telecom industry. Early arrivals have a lot to gain, and those who are left behind have a lot to worry about

Since June 2023, Drakshi has been writing articles of Artificial Intelligence for govindhtech. She was a postgraduate in business administration. She was an enthusiast of Artificial Intelligence.


Please enter your comment!
Please enter your name here

Recent Posts

Popular Post Would you like to receive notifications on latest updates? No Yes