Network Functions Virtualization Infrastructure

Description

The Network Functions Virtualization Infrastructure (NFVI) is the foundational layer for virtualized and cloud-native network architectures as defined by ETSI and adopted by 3GPP. It comprises the totality of all hardware and software components that collectively create the environment in which Virtualized Network Functions (VNFs) are deployed, managed, and executed. Conceptually, it abstracts physical resources into a pool of virtual resources, decoupling network software from proprietary hardware. This abstraction is crucial for enabling the elasticity, scalability, and automation promised by network virtualization.

The architecture of NFVI is typically structured into three domains: the compute domain, the storage domain, and the networking domain. The compute domain consists of commercial off-the-shelf (COTS) servers that provide processing power via virtual machines (VMs) or containers. The storage domain includes both local and shared storage systems, virtualized to provide block, object, or file storage to VNFs. The networking domain encompasses the physical network interfaces, switches, and routers, along with their virtualized counterparts like virtual switches (e.g., Open vSwitch) and virtual networks, which provide connectivity between VNFs and to external networks.

At the software heart of the NFVI lies the virtualization layer, most commonly a hypervisor (for VM-based deployments) or a container runtime (for container-based deployments). This layer is responsible for abstracting the physical resources, partitioning them, and presenting them as virtual resources to the VNFs. It manages the lifecycle of these virtual resources, ensuring isolation, security, and performance for each VNF. Above this, a Virtual Infrastructure Manager (VIM), such as OpenStack or Kubernetes, provides the orchestration and control plane for these virtual resources. The VIM handles tasks like resource inventory, allocation, placement, and monitoring, interfacing with higher-level management and orchestration systems like the NFVO.

The role of NFVI within the 3GPP ecosystem is to host core network functions (like the AMF, SMF, UPF) and other network applications. By providing a standardized, vendor-agnostic platform, it allows operators to deploy network functions from different vendors on a common infrastructure, breaking vendor lock-in. It enables rapid service deployment through automation, efficient resource utilization via pooling and sharing, and scalable operations that can elastically grow or shrink based on demand. The performance, reliability, and security of the NFVI are therefore paramount, as they directly impact the service quality of the entire mobile network.

Purpose & Motivation

NFVI was created to address the limitations of traditional telecom networks built on proprietary, integrated hardware appliances. These legacy systems were costly, slow to deploy and upgrade, and led to vendor lock-in, stifling innovation and increasing operational complexity. The primary motivation was to leverage IT virtualization technologies and cloud principles to transform the telecom infrastructure, making it more agile, cost-effective, and scalable.

The historical context is the industry-wide shift towards Network Functions Virtualization (NFV), championed by ETSI. NFVI is the realization of the NFV vision's infrastructure pillar. It solves the problem of hardware dependency by abstracting network functions into software that can run on standard high-volume servers, storage, and switches. This shift allows network operators to benefit from the economies of scale and rapid innovation cycle of the IT industry.

Furthermore, NFVI enables the dynamic orchestration and automation of network services, which is essential for supporting modern demands like network slicing, edge computing, and on-demand 5G services. It provides the foundational agility required to compete with cloud-native service providers and to efficiently manage the massive scale and diverse service requirements of 5G and beyond.

Key Features

• Abstraction of physical compute, storage, and networking resources
• Support for both Virtual Machine (VM) and container-based workloads
• Integration with Virtual Infrastructure Managers (VIMs) like OpenStack/Kubernetes
• Enables multi-tenancy and isolation between different VNFs
• Provides APIs for automated resource lifecycle management
• Forms the basis for cloud-native network function deployment

Evolution Across Releases

Defining Specifications