Traffic Collection, Contextual Analysis, Data-driven Optimization for 5G
IMDEA Networks Institute participates as partner without assigned budget
  • Financed by: CoCo5G project received funding from ANR, but IMDEA Networks participation is self-funded. ANR-22-CE25-0016
  • Duration: October 2022 to March 2026
  • Contact: Marco FIORE, Principal Investigator for IMDEA Networks
Visit webpage arrow_right_alt

In 2021, 5G started being deployed worldwide, with minimum coverage that is expected to accelerate and reach a worldwide technology share above 20% by 2025. In France, 5G traffic currently represents less than 1% of the total cellular network traffic, but a major deployment effort is expected before the 2024 Olympics. This shall favor an earlier adoption of the technology with respect to other European countries, making France an ideal national-scale case study for monitoring and understanding how the 5G evolution will impact our lives.

Most of the current 5G system deployments, both in France and globally, are Non-Stand-Alone (NSA): the access technology is the 5G New Radio (NR), which however still relies on the previous-generation 4G cellular core network. Coexistence between 4G and 5G is indeed expected to last many years, with successive NSA architecture variations leading to the gradual integration of the 5G Core (5GC) control and data planes. Notably, 5GC shall be much more pervasive than its 4G counterpart, in particular to favor a widespread multi-access edge computing (MEC) infrastructure and support low-latency services. MEC traffic local-breakout points shall be planned at or even beyond the legacy points-of-presence (PoP) of the network operator: as an example, major national mobile access providers in France rely today on around a dozen 4G core sites, while in 5G this number is expected to be increased by one order of magnitude at least, likely more.

A 5GC composed of an ubiquitous MEC infrastructure that is highly softwarized paves the way to unprecedented mobile network performance. Among the most prominent features, proximity to the end user will enable applications requiring Ultra-Reliable and Low Latency Communication (URLLC); and, the possibility to dynamically configure MEC network functions will support an automated management of MEC resources that is efficient and robust to anomalous fluctuations in the mobile traffic. Ultimately, 5GC promises to facilitate the emergence of whole new families of innovative and life-changing mobile services.

Nonetheless, significant uncertainties threaten the timeline and the very same progress towards such full-fledged 5G deployments. First, URLLC-based and other MEC-enabled applications represent a brand-new market whose growth and economic worth are hard to predict: although the first tangible examples of commercial millisecond-response-scale edge services are appearing, whether they will warrant the huge capital expenditure to deploy a MEC infrastructure is still unclear. Second, the additional complexity of orchestrating network functions and resources in a MEC environment calls for zero-touch management solutions, whose algorithmic design and performance evaluation in presence of 5G data traffic are open research problems. Third, the sustainability of 5GC in terms of energy consumption is a major societal concern, and quantitative assessments from early deployments as well as dedicated energy-prudent MEC solutions are needed to dispel all doubts in that direction.

The main reasons for these uncertainties in the deployment of the full and long-term 5G stack are the lack of experience on the promised novel advanced services brought by 5G. The bit rate increase of the most conventional traffic type is the major characteristic beyond the eMBB (enhanced Mobile Broadband) class, designed as an extension of the default 4G class, particularly useful for interactions between mobile users and Content Delivery Network (CDN). Besides the eMBB conventional class, two brand-new classes are in the process of further standardization, such as the URLLC class.

The CoCo5G project aims at addressing these open questions, leveraging on network data analysis from commercial 5G networks, during the very first year of adoption of 5G in France and Europe.

CoCo5G Objectives and research hypothesis

We propose to carry out a comprehensive set of data-driven analyses that will close the above gaps, and relieve the uncertainty about the interest, viability and advantages of an architectural migration from a NSA 5G model to a combined MEC-5G system. To this end, we will take advantage of the anticipated 5G deployments in major conurbations of France, also stimulated by the 2024 Olympic Games preparation, and perform activities that span from data collection to 5G network functionality assessment. Specifically, our set of precise, measurable and logically connected objectives is as follows:

• Objective 1: Collection of novel measurements dataset combining 4G and 5G data traffic. We will collect a first-of-its-kind longitudinal nationwide dataset of combined 5G and 4G data traffic that will enable unprecedented analyses of 5G network evolution and design of 5G network solutions. The construction of this dataset is essential, considering the lack of such datasets due to the recent 5G deployments in France.

• Objective 2: Extensive analysis of the evolution of in France and the dynamics of 5G traffic for various mobile services usages. We will develop a structured knowledge base about how 5G will be changing mobile service usages over time, and across diverse user groups characterized, e.g., by different socioeconomic statuses. The recent development of 5G networks comes with various data analytics studies, which are fundamental for the understanding of novel usages of services and the design of network management processes.

• Objective 3: Evaluation of existing analytics for classification, prediction and anomaly de- tection within real-world high-detail per-service mobile network data, and tailoring them to the specifications of the management of resources at different network levels. We will evaluate a range of decision-making models proposed for mobile network management, and assess their effectiveness against real-world 5G data traffic across the heterogeneous usages and user groups identified before; we will improve the models to overcome their identified deficiencies.

• Objective 4: Demonstration of the integration of data analytics within next-generation cognitive network architectures in three practical case studies. It is critical to deploy the designed models of traffic analysis and network optimizations on a variety of practical use-cases that address key 5GC network functions. We will apply the refined models above to three practical case studies, i.e., (i) energy-prudent 5G NR control, (ii) URLLC service support, and (iii) automated anomaly response at MEC, demonstrating the viability of 5GC solutions in realistic data traffic settings and contributing to the design of closed-loop automation systems.

Contract nb: ANR-22-CE25-0016