Very high throughputs and connection densities are critically important in “megacities”. In the near future, these areas will also include large quantities of wireless connected machine-type communications (MTC) devices.
To tackle challenges like the high degree of intermixed services, 5G will need to efficiently deliver the expected Quality of Service (QoS) and Experience (QoE) to all such services. Classic Key Performance Indicators (KPIs) are not enough to optimise and grasp the network status. This is where Key Quality Indicators (KQIs) come into play for service-oriented analysis and metrics, in the management of 5G networks. However, the continuous gathering of KQIs has its own challenges: the use of secure HTTP and high-layer encrypted protocols limit the traffic inspection. Access to application layer KQIs is also very limited in both ends of the communication as the application user-experience is generally out of reach of the cellular management monitoring.
Goals of the trial and testbed usage
The ONE5G megacity trial focused on innovative end-to-end (E2E) network management for 5G infrastructures using KQI-based monitoring and characterisation of E2E performance.
Demonstrating enhanced QoE metrics in terms of achieving a proper assessment of the network status by translating low-level indicators into higher layer performance metrics.
Target KPI: end-to-end latency and packet loss rate and the efficiency of the slice negotiation process.
Validating innovations for E2E monitoring schemes based on actual user QoE as enablers for future network management and optimisation solutions in megacity environments: integrated in a full indoor LTE network deployed at Malaga University.
Validating slice negotiation and management functionalities in megacities: implemented in WINGS ICT Solutions testbed.
Architecture and Trial Overview
The high-level architecture includes 5G technologies focusing on eMBB megacity services.
Features of the trial include:
5G infrastructure emulating a megacity environment with a set of small cells.
Software components boosting the infrastructure with 5G functionalities.
5G KPI to KQI modelling component for managing the megacity infrastructure. KPI/KQI monitoring and estimation.
5G slice negotiation and management component for automated network slice negotiation between the vertical side (requirements) and operator side (capabilities and availability) and for forwarding decisions towards the megacity infrastructure.
Innovative E2E network management for 5G using KQI monitoring to achieve a proper assessment of the network status by translating low-level indicators into higher layer performance metrics alongside innovative network slice negotiation and management techniques.
Implementation of tools needed for predicting KQIs based on cellular low-layer performance metrics and configuration parameters. The two related goals are: showing KPIs and associated service KQIs under different circumstances, comparing estimated and directly measured KQI values. A baseline network optimisation to show the capabilities of this approach.
Implementation of mechanisms for the automated negotiation of price offers for certain service quality levels for megacity scenarios with price levels covering environment heterogeneity, variable service/traffic demand (e.g. accommodating eMBB, URLLC, mMTC) and network aspects (e.g. network utilisation etc.).
Full indoor LTE network accessible through a REST API, a testing UE and remote client, running a mapping script. This script translates in real time low layer metrics collected from the network into service KQIs.
Outcomes and Impacts
A live prediction graph showing if the estimated “a priori” E2E latency and packet loss rate are above a certain predefined threshold. It also shows values measured “a posteriori”, demonstrating the prediction capability of the system.
Demonstrating the negotiation process between the operator and vertical offered prices. The two prices may or may not converge to an agreement after counter-offers (service offer with lower qoE/QoS or shorter time duration).
Partners involved: B-COM, Malaga University, WINGS ICT Solutions
Other Partners: Fraunhofer Heinrich Hertz Institute (Coordinator), Aalborg University, Centre National de la Recherche Scientifique, Freie Universität Berlin, Huawei Technologies Duesseldorf GmbH, Intel Germany GmbH, NOKIA Bells Lab (Denmark, Germany), Orange, Samsung Electronics UK, Telefónica