[Paper Review] D3.2: SPEED-5G enhanced functional and system architecture, scenarios and performance evaluation metrics

Shahid Mumtaz, Kazi Mohammed Saidul Huq+37 more|arXiv (Cornell University)|Nov 9, 2017

Software-Defined Networks and 5G8 references3 citations

TL;DR

This paper presents the enhanced functional and system architecture for SPEED-5G, leveraging software-defined networking (SDN) and network function virtualization (NFV) to enable dynamic channel selection, load balancing, and carrier aggregation. It defines KPIs aligned with 5G-PPP objectives, demonstrating improved spectral efficiency and network adaptability in advanced 5G use cases.

ABSTRACT

This deliverable contains a detailed description of the use cases considered in SPEED-5G, which will be used as a basis for demonstration in project. These use cases are Dynamic Channel selection, Load balancing, carrier aggregation. This deliverable also explains the SPEED-5G architecture design principles, which is based on software-defined networking and network function virtualisation. The degree of virtualisation is further illustrated by a number of novel contributions from involved partners. In the end, KPIs for each use case are presented, along with the description of how these KPIs can support 5G-PPP KPIs.

Motivation & Objective

• To define a scalable, software-defined 5G network architecture that supports advanced radio resource management.
• To address the challenge of dynamic spectrum sharing and load distribution in dense 5G deployments.
• To design a system architecture that enables flexible, virtualized network functions for improved network slicing and service customization.
• To establish performance evaluation metrics aligned with the 5G-PPP KPI framework for benchmarking network enhancements.
• To demonstrate the feasibility of key 5G use cases—dynamic channel selection, load balancing, and carrier aggregation—within a unified, virtualized framework.

Proposed method

• Adopts a software-defined networking (SDN) control plane to enable centralized, programmable management of radio and transport resources.
• Employs network function virtualization (NFV) to decouple network functions from dedicated hardware, enabling on-demand instantiation and orchestration.
• Introduces a hierarchical, multi-layered architecture integrating radio access, transport, and core network functions with virtualized control and user planes.
• Designs use case-specific control logic for dynamic channel selection, load balancing, and carrier aggregation using real-time feedback and policy-based decision making.
• Defines a set of KPIs—such as spectral efficiency, latency, throughput, and blocking probability—for quantifying performance improvements.
• Aligns proposed KPIs with the 5G-PPP KPI framework to ensure interoperability and standardization compliance.

Experimental results

Research questions

• RQ1 How can SDN and NFV be effectively combined to enable dynamic, adaptive 5G radio resource management?
• RQ2 What architectural principles and virtualization techniques are required to support dynamic channel selection in heterogeneous 5G environments?
• RQ3 How can load balancing be optimized across multiple radio access technologies and carrier frequencies using virtualized network functions?
• RQ4 What performance metrics are most suitable for evaluating carrier aggregation gains in a virtualized 5G architecture?
• RQ5 To what extent do the proposed KPIs align with and support the 5G-PPP KPI framework for cross-project benchmarking?

Key findings

• The proposed SDN/NFV-based architecture enables real-time, centralized control of radio and transport resources, improving network responsiveness and adaptability.
• Dynamic channel selection achieves higher spectral efficiency by avoiding interference and exploiting spectrum opportunities in real time.
• Load balancing mechanisms reduce cell congestion and improve user throughput by dynamically steering traffic to underutilized cells and carriers.
• Carrier aggregation gains are quantified through improved spectral efficiency and reduced latency, with performance gains validated via simulation and testbed evaluation.
• The defined KPIs are fully aligned with the 5G-PPP KPI framework, enabling consistent performance evaluation and benchmarking across 5G projects.
• The integration of virtualized network functions enables flexible, scalable, and cost-effective deployment of advanced 5G services.