AI by design - to provide data driven operations of networks
Network complexity is rapidly increasing, and communication service providers must find a way to keep up. Here, we explain why AI by design and zero-touch networks could be the answer.
In a previous blog post, we outlined a fast-changing market with new service demands requiring new business models. Communication service providers must adapt to address even more complex use cases and the new opportunities enabled by artificial intelligence (AI).
AI adoption for network orchestration, management and operations is extremely vital. It helps address the complexity of today’s networks, drive efficiencies and improve customer experience as well as open new revenue streams.
What is a zero-touch network?
As new technologies like 5G, IoT and Cloud are gaining traction, networks are becoming more complex. It is becoming challenging for humans to keep up with this complexity and keeping networks operating at the level prescribed by the evolving services. We are already flooded by data and this is only the beginning. Running a 5G network, including data points in IoT, combined with demands of mission critical use cases is only possible when applying Artificial Intelligence (AI), automation and data analytics to drive “Data Driven Operations” of telecom networks.
The Data driven operations help us move closer to our vision of “Zero-touch networks”.
The zero-touch network is a simplified and software-based network platform that utilizes automation and AI to augment human expertise in its operations. Technologies and capabilities are required to work in harmony to achieve the vision of zero-touch networks. A zero-touch network is based on AI, which means that network components and operational capabilities needs to be implemented to support AI technology. Products and services used in the system need to be built with AI incorporated from the start. A number of these technologies and mechanisms are required to achieve the vision of zero touch, such as:
- Cloud native architecture
- AI-ready data collection and management
- Closed-loop automation
- Consolidated data analytics, management and orchestration
- AI operations
With more network functions decoupling the hardware from the software through virtualization technologies, service providers will experience challenges related to change management, standardization of data, service models and APIs, as well as the lack of interoperability across their supply chain. Therefore, it is recommended to implement a stepwise approach to the zero-touch network, focusing on solving one specific problem at a time, resulting in less risks and more time to learn. With this approach, you can also introduce changes quickly as the organization learns. This is the foundation of zero-touch and will reduce costs while at the same time introducing new services and network capabilities to improve customer experience.
AI by design
So what is our approach to AI? Ericsson’s technology strategy foresees the adoption of AI across the product and services portfolio in all parts of the network architecture. The purpose is to empower humans and machines to transform an engineered and adaptive network into a continuous learning network which better serves our customer needs. We use AI to:
- Boost automation capabilities within operations and services
- Power Ericsson’s solutions
- Create new business opportunities in telecom and IoT
We also believe that AI should be applied where it solves specific challenges for telecom service providers – creating value where it matters most. It’s not a general-purpose tool intended for any use case imaginable. Our AI solutions are embedded throughout the network, built by people with extensive AI and telecoms expertise and with an AI-first mindset for every product or service. This is why we call it AI by Design: AI designed to solve the telecom challenges of today and tomorrow.
To meet these challenges, we have focused our work with AI around the problems that we think really matters, such as:
- Maximizing the return on investment on installed base without the need to visit sites or adding new hardware
- Securing always on, high-performing networks for consumers
- Delivering business critical enterprise applications and critical infrastructure for national services
- Living up to increased consumer expectations and a flawless QoS experience for industries
- Realize the value of cloud with improved agility, lowered cost and building the foundation for new services
- Handling higher demand while keeping the operational costs at a minimum
- Building flexible networks following actual usage and ensuring energy consumption is kept at a minimum while maintaining the right performance levels
Read the OVUM report "Achieving a true zero-touch network vision" for more informationDownload report
Technology building blocks for zero-touch networks
All components of your system and the operations processes needs to be implemented with AI integrated from start. This is what we call AI by design. This requires AI to be implemented in (see figure 2):
- BSS through customer intelligence support
- A management and orchestration solution including automated insight operations, autonomic incident management, predictive analytics for assurance, self-organizing networks and advanced workload placement
- Radio access with AI-enabled handover, 5G-aware traffic management, evolved load balancing at release, uplink-traffic triggered mobility, traffic-aware carrier aggregation and augmented MIMO sleep
- Cloud core with machine learning assisted paging
- Ericsson services including smart design, proactive customer support, network service operations, advanced optimization and cloud and IT service operations
Automation is an essential part of the solution. Software that is able to operate autonomously and make smart decisions in a complex environment is referred to as an intelligent agent (a practical implementation of artificial intelligence and machine learning).
With the introduction of Ericsson Operations Engine combined with Network automation solutions, by means of products, processes and tools, we accelerated our journey from manual, reactive, incident-driven operations, towards sero-touch networks and truly proactive, data-driven operations
Forward-looking network operators and digital service providers require an automated network and business environment that can support them in the transition to a new market reality characterized by 5G, the Internet of Things, virtual network functions and software-defined networks. Conventional machine learning (ML) models provide many benefits to mobile network operators, particularly in terms of ensuring consistent quality of experience. However, on top of creating a substantial network footprint, the large data transfer required by conventional ML models can be problematic from a data privacy perspective. Federated learning (FL) makes it possible to overcome these challenges by bringing “the computation to the data” instead of transferring “the data to the computation.
Read more about how to migrate from a conventional ML model to an FL model and significantly reduce the amount of information that is exchanged between different parts of the network, enhancing privacy without negatively impacting accuracy in this Ericsson technology review article: Privacy-aware machine learning with low network footprintDownload
The concept of “cognitive intent-based management” using advanced AI technologies is gaining traction. Service providers and their enterprise customers will specify “what” behavior they want from the network and the system then decides “how” to fulfil this request using closed-loop automation, with minimal human intervention whenever possible. We are investing heavily to incorporate the machine reasoning capabilities to enable these value propositions in our product and service offerings.
Conclusion and further reading
A zero-touch network enabled by a software-based network platform, automation and AI will minimize the need for human intervention in its operations. This requires all components of your system, including products and operational processes to be “AI by Design”.
In next blog post we will present an interesting customer use case of cognitive OSS. Stay tuned!