Strategy Analytics and P.I. Works recently published a white paper on the impact of AI and automation on 5G network management. This white paper provides valuable insights on how service providers looking to deploy and operate complex, scalable and converged pre-5G and 5G networks can leverage next generation automation and AI capabilities from today. Below is a summary of the key findings. To download the full white paper please visit: https://pi.works/5G_AI
5G is the first mobile technology that not only enhances mobile broadband use cases, but also natively supports massive Machine Type Communication (mMTC) and Ultra-Reliable Low-Latency Communication (URLLC). The diversity of mobile services and their distinct requirements position 5G as an essential infrastructure for our future connected society. But the resultant use cases also demand very high network performance and place challenging requirements on 5G network management and service delivery.
Coexistence of multiple radio technologies, such as 4G LTE, 5G SA/NSA along with legacy 2G, 3G, within the same geographic area, combined with use of low, medium and high frequency bands (including carrier aggregation) results in a huge network complexity. Delivery of network design and harmonious optimization across different bands, technologies and network architectures must be handled centrally. It must also factor in various dynamic parameters such as traffic demand, network load and user behavior.
Yet, while the 5G network will be much more complex than 4G, operators must deliver the new 5G services at the same or even lower cost than 4G in order to keep their business sustainable. There are three key areas where operators must leverage the benefits of automation and AI to make 5G both feasible and profitable.
- Optimize 5G efficiency across vast numbers of small / macro cells in a massively densified network as well as across multiple diverse use cases (eMBB, URLLC, mMTC)
- Fully automate network performance and operational processes end-to-end (E2E)
- Lower overall network management costs
Slice Aware Optimization must be automated to match QoS requirements for diverse network slices. Based on the information from both physical and virtual network functions, network automation can dynamically manage physical resources to optimize different slices. P.I. Works and Turkcell recently initiated a research project to automate optimization of Turkcell’s network slices. With service-aware dynamic slice management, Turkcell aims to enable more flexible use of network resources and guarantee customer SLAs.
Backhaul Aware Optimization needs to be an integral part of E2E 5G optimization. E2E traffic load must be balanced across both the access and backhaul infrastructure. Real time monitoring of overutilized and underutilized transport links, as well as the prediction of backhaul traffic using AI methods deliver key savings for mobile operators. A recent trial of P.I. Works solution with a Tier 1 European Operator with approximately 20 million subscribers generated $161.3 thousand of savings per year during the trial in a cluster with 272 leased backhaul links and ensured improved service quality. This outcome corresponds to extrapolated savings of $1.52 million per year for a network-wide deployment.
Smart Capacity Planning uses predictive analytics and traffic models to recommend capacity changes. Today this is often limited to individual cells. As operators move to more densified and heterogenous network topologies, they need to look at statistics for groups of cells i.e. ‘cell clusters’ in order to dramatically improve capacity management across large numbers of small cells in a metro area.
Predictive Energy Saving utilizes techniques based on data collected from live network usage, e.g., the active numbers of user equipment (UEs) associated with a base station in a given time period and PRB utilization trend. These algorithms are able to make very precise predictions. Such energy saving projects were recently implemented by P.I. Works in number of operators with multi-vendor, multi-technology networks. The results of these projects indicated strong double-digit savings in the energy consumption.
Automated Advanced MIMO Optimization will be critical to enhancing spectrum optimization through reducing pilot contamination of advanced MIMO deployments in pre-5G and 5G networks. With the introduction of massive MIMO and 3D beamforming as part of 5G NR, the automated management of predefined coverage configurations based on traffic predictions, QoS requirements, traffic type, load per cell or per transport link will help drive even greater efficiency and enhance end user experience.
Wide Area Resource Coordination improves the service reliability particularly for the URLLC use cases in 5G and orchestrates all the network elements. As centralized network automation functions collect data and forecast traffic across cell clusters and across the metro area, network resources can all be coordinated over a wider area.
Ron Robinson, Vice President P.I. Works Americas said “Mobile operators will require network automation and AI capabilities to become ‘5G ready’. But many of these approaches can be applied to LTE and LTE Advanced Pro to capture many technical, business and process transformation benefits today. Early adoption of advanced automation and AI technologies prepare operators for 5G, which is an important milestone in reaching Zero-Touch Network Orchestration. P.I. Works is the trusted partner of operators on the road to 5G and helps them achieve their network transformation goals.”
To learn more about P.I. Works and Strategy Analytics please visit:
P.I. Works also plays an important role in the development of key industry standards that define the future of mobile networks. P.I. Works actively contributes to the European Telecommunications Standards Institute (ETSI), 3rd Generation Partnership Project (3GPP) standardization forum, Global TD-LTE initiative (GTI) and open source initiatives.
Sanayi Mah. Teknopark Bul.
No:1/3A 101 34906 Pendik/Istanbul
T: + 90 216 999 1099
F: + 90 216 912 1099