AI Networking and Tools for 5G and Beyond

- Overview

Artificial intelligence (AI) and 5G are synergistic ingredients that can fuel future innovations. AI can help improve 5G system performance and efficiency. 5G can support AI deployment and enable the development of new AI uses with distributed AI.

AI can improve 5G wireless capabilities in several ways, including: 

• Optimize network performance: AI algorithms can analyze data generated by the network to provide insights into network usage patterns, traffic congestion, and performance issues.
• Optimize energy consumption: AI can analyze network traffic and demand patterns to dynamically adjust power usage and optimize resource allocation. This can lead to significant energy savings and reduced operational costs for network operators.
• Beamforming: A machine learned algorithm can assist the 5G cell site to compute a set of candidate beams, originating either from the serving or its neighboring cell site.
• Predictive maintenance: AI can empower software tools to take quick actions and respond immediately enabling predictive maintenance. 
• Process data: AI can allow for the processing and analysis of the massive amounts of data generated by IoT devices. 
• Improve AI-enhanced experiences: Using 5G with AI can allow for distributed AI processing that would offer more flexibility for new functions.

AI/ML is being used across 5G system, including management and orchestration, core network, and RAN.

- Bringing AI To Wireless Networking

In the world of wireless networking, AI is already showing enormous value. Machine learning and neural networks simplify operations, expedite troubleshooting, and provide unprecedented visibility into the user experience. But we are just on the cusp of its true potential.

Emerging applications require wireless connectivity with tremendously increased data rates, substantially reduced latency, and growing support for a large number of devices. These requirements pose new challenges that can no longer be efficiently addressed by conventional approaches.

AI is considered as one of the most promising solutions to improve the performance and robustness of 5G and B5G (Beyond 5G) systems, fueled by the massive amount of data generated in 5G and B5G networks and the availability of powerful data processing fabrics. 

As a consequence, a plethora of research on AI-based communication technologies has emerged recently, promising higher data rates and improved QoS with affordable implementation overhead.  

With the continuous improvement of networks, the extent of new applications is limited only by our imagination. Network planning, construction, maintenance, optimization and operations need to evolve intelligently to support new services, applications and user experiences. AI needed to be introduced to increase the accuracy of user experience-based network planning to make it more agile, and to quickly and more intelligently resolve user experience problems.

- Agile Network Construction

AI is applied to each phase of 5G network construction to make network planning more accurate and the rollout more efficient. Multiple data points from areas covering the 5G business, users and the evolution of existing technology are used with machine learning and iterative computing to quickly and accurately create plans for different scenarios. 

Technologies such as photogrammetry, optical character recognition (OCR), voice recognition and computer vision are introduced in the survey, design, commissioning, integration and acceptance phases to continuously improve engineering automation and quality of delivery.

[Columbus Circle, New York City - Civil Engineering Discoveries]

- Intelligent O&M

The co-existence of 2G, 3G, 4G and 5G networks drastically increases the number of connections between people and things, while providing a wide range of services for users. Consequently, the number of service requests and problems faced by operations and maintenance (O&M) personnel also increases. 

According to data analysis over the past few years, network O&M problems are increasing by 5 per cent annually. Conventional O&M practice cannot be sustained, as people are the main workforce, with just isolated O&M tools for assistance. The introduction of AI for man-machine collaborative IT operations (AIOps) is becoming an inevitable trend. 

AIOps O&M will not mean breaking the O&M system and abandoning existing tools. Instead, an O&M knowledge platform using existing models is added, which can drive the evolution of existing domain- and phase-based, human-dominant O&M to man-machine collaborative O&M.

AIOps will not substitute people but will enable them to play a greater role with the assistance of machines. O&M talent will take new positions such as network policy engineers, orchestration engineers and data analysts. People will focus on more important roles in intent design, troubleshooting and key decision-making. 

- Smart Operations

5G is much more than a wireless networking technology evolution -- 5G is driving the telecom industry to integrate and interoperate in revolutionary ways, and new business and deployment models are needed. 5G is set to become a core element of the global digital economy. Yet, the features that make 5G attractive, such as a shared infrastructure, also make it a security risk. To experience 5G's many advantages, organizations will need to assess their own risk and protect their networks and data accordingly.

5G is ushering in a new era of communications. It will bring better services, applications and unprecedented experiences to consumers. It will also create an opportunity for operators to break the conventional "pipe" business model, enabling them to develop new digital services, explore new business models and foster new industry partnerships.

- Open Ecosystems

New 5G services and applications will likely require much more bandwidth, but subscribers -- both enterprises and consumers -- won’t necessarily be willing to pay more for service. Hence, communications service providers (CSPs) must be able to curtail costs while building a network that scales. The traditional model of building networks with very expensive, single network, purpose-built platforms from a single vendor that locks out competition may no longer be sustainable in the 5G era. 

The enterprise realm and data centers have already embraced a new model: leveraging cloud technologies such as software-defined networking (SDN) and network functions virtualization (NFV) based on open principles. By leveraging software-centric, open-reference architectures, CSPs can achieve cloud economics and agility while enabling a multivendor ecosystem, reducing complexity and lowering costs for their 5G networks.  

To a large degree, new open architecture business models built on open source can contribute to the reduction of risk for 5G network rollouts given common frameworks and tools. This shift to open may also equate to more vendor choice, as challengers to incumbent “tech giants” will have greater access to ecosystems and even the largest CSPs. For 5G to reach the potential of its promise, core elements in the network must evolve to support mobile broadband increasing requirements for mission-critical, ultra-low-latency applications.

[More to come ...]