Optimizing 5G network slicing: An end-to-end approach with isolation principles

5G Standalone (SA) networks introduce a range of new applications, including enhanced Mobile Broadband (eMBB), Ultra-Reliable Low-Latency Communication (URLLC), and massive Machine-Type Communications (mMTC). Each of these

applications has distinct network requirements, which current commercial network architectures, such as 4G and 5G Non-Standalone (NSA), struggle to meet simultaneously due to their one-size-fits-all design. The 5G SA architecture addresses this challenge through Network Slicing, creating multiple isolated virtual networks on top a single physical infrastructure. Isolation between slices is crucial for performance, security, and reliability. Each slice owns virtual resources, based on the physical resources (e.g., CPU, memory, antennas, and network interfaces) shared by the overall infrastructure. To deploy Network Slicing, it is essential to understand the concept of isolation. The Third Generation Partnership Project (3GPP) is standardizing security

for Network Slicing, focusing on authentication, authorization, and slice management. However, the standards do not clearly define the meaning of isolation and its implementation in the infrastructure layer. In this paper, we define and showcase a real-life Proof of Concept (PoC), which guarantees isolation between slices in 5G

SA networks, for each network domain i.e., Radio Access Network (RAN), Transport Network (TN), and 5G Core (5GC) network. Furthermore, we describe the 5G SA architecture of the PoC, explaining the isolation concepts within the Network Slicing

framework, how to implement isolation in each network domain, and how to evaluate it.