Unleash high-performance virtual networks

Network functions virtualization (NFV)

Network functions virtualization (NFV) takes modern networks one step closer to the cloud, making the deployment and management of network functions, services and applications completely virtual.

Network functions virtualization (NFV) is a modern network architecture approach that decouples the network’s various applications from its hardware resources including compute, storage and other network hardware. This creates a virtualized layer of network functions deployed as virtual machines (VMs) or containers that can be shared by all network applications.

Using virtual, software-based applications where dedicated physical boxes once stood, NFV makes it possible for communication service providers (CSPs) to manage, orchestrate, and expand network capabilities on-demand, wherever they’re required across the network.

Most of the world’s 5G networks run today on NFV infrastructure. As a precursor to the ongoing cloud native revolution, expect both NFV- and cloud native infrastructures to co-exist for many years to come.

From hardware to the cloud: the story of network function evolution

Just as technologies and network demands have evolved over time, so too have network functions. From first generation networks to today’s 5G networks, network functions have evolved from being an element tightly coupled to the network’s hardware to one that is virtualized, containerized, and now cloud native.

1979- Proprietary hardware and software

For many decades, from 1G to 3G networks, mobile networks have relied on physical network functions, deployed either on purpose-built, proprietary hardware as a network element, or on software running on dedicated network hardware.

2013- Networks go virtual

In 2013, the standards organization ETSI introduced the very first specifications for the virtualization of network functions, driven by demands from the industry to make networks more vendor agnostic, agile and cost efficient. This new architectural concept decoupled the network’s functions from its hardware, making it possible to run virtual network functions (VNFs) on virtual machines hosted on general-purpose servers.

The breakthrough of software-defined networking around the same time meant that complex, virtualized network environments could be automated and orchestrated with speed and agility. For the first time, new network functions could be deployed and scaled almost instantly.

2017- The era of cloud native

Cloud native is the next step in the deployment evolution of network functions, advancing both hardware and software decoupling, lifecycle management as well as the decomposition of software and lifecycle management and making it possible to deploy cloud native applications on bare metal infrastructure using containers and microservices, rather than virtual machines and a virtualization layer. The cloud native paradigm shift, which ETSI began to integrate into NFV standards around 2017, has become the cornerstone of new telecom operational models for both 5G and future 6G networks. Benefits include higher resource utilization, faster lead times for new applications, and TCO savings.

Ready to take the next step from virtual to the cloud?

Skip to cloud native

Benefits of NFV: making networks more efficient, agile and innovative

By making networks more resource efficient, more performant and much more agile in launching, scaling and managing new services, NFV makes it possible for CSPs to meet and stay ahead of the growing demands placed on today’s high-performance networks.

Lower day-to-day costs

The shift away from purpose-built hardware towards general-purpose servers lowers both the cost of outlay (CAPEX) and maintenance (OPEX). This is complemented by additional automation-based cost savings, as well as lower overall network energy consumption.

Improved performance

Through dynamic resource allocation, NFV makes it possible to upscale and downscale network functions based on demand. This makes it possible for CSPs to improve load balance and meet traffic spikes without overprovisioning.

Deliver services faster

With vendor-agnostic multi-tenant support, multiple virtual and cloud-native network functions can be aggregated onto a single platform, meaning that multiple services and applications can be served and scaled simultaneously.

Innovation on-demand

Shifting from physical to virtual infrastructure significantly improves the speed and agility with which network functions and services can be launched, updated or sunsetted. Together with faster and more agile development cycles, CSPs are better equipped to test new services and respond to market demands in no time.

NFV architecture: the basics

The NFV infrastructure (NFVI) platform sits at the heart of NFV architecture. By enabling and orchestrating the efficient, flexible and scalable deployment of virtual network functions (VNFs), the NFVI is the backbone that makes today’s high-performance networks possible.

NFVI typically consists of many architectural building blocks, including:

a hardware layer of compute, network and storage resources
virtualized resources, such as VNFs or cloud-native container-based microservices
a virtualized execution layer that provides orchestration and run time environment for VNFs, containers or cloud-native functions
a management, automation and network orchestration layer

Above: Multi virtual network function (VNF) NFV infrastructure that comprises a hybrid infrastructure-as-a-service and container-as-a-service (CaaS) platform approach, serving both VNFs and cloud-native network functions (CNFs)

The network cloud space is evolving rapidly. Get up to speed with some of the key NFV and cloud technologies and terms below.

Virtual network functions (VNFs) vs. cloud-native network functions (CNFs)

VNFs are network functions deployed in virtual machines (VMs) on a hypervisor layer in a traditional virtualized network architecture. In contrast, CNFs are fully cloud-native network functions deployed in containers within a simplified microservices-based architecture, typically using Kubernetes as a container orchestration platform. This makes CNFs more agile, flexible and highly scalable.

Both VNFs and CNFs can run in parallel on the same NFVI infrastructure by deploying a container-as-a-service (CaaS) platform on top of the exiting virtual infrastructure manager (VIM). This is an effective alternative for CSPs that already deploy performant horizontal cloud platforms with multiple VNFs, yet are looking to advance their cloud transformation.

Cloud infrastructure

Cloud infrastructure architecture makes it possible for various applications to share compute, network and storage resources for their needs, reducing time to market for new services and making it easier to manage them with shorter lifespan.

Cloud infrastructure architecture is typically delivered to users “as a service”. While Infrastructure as a Service (IaaS) has been the most common way of using cloud infrastructure architecture to deploy applications, Containers as a Services (CaaS) will become the primary deployment method under cloud native deployments. Learn more about cloud infrastructure.

Software-defined networking

Software-defined networking (SDN) in networks emerged around the same time as network virtualization as a means to achieve the necessary automation and orchestration to manage complex, virtualized network environments. Separating the network’s control plane from the data plane, SDN enabled the centralization of network control, meaning network functions could be programmed, managed more dynamically, and provisioned as a service . While it continues to be important, it is not typically the primary focus in discussions about modern network architecture, which have largely moved on to cloud-native solutions. SDN today is often built in as a function in the NFVi solution rather than a separate domain.

Network function virtualization infrastructure (NFVi)

Network function virtualization infrastructure (NFVI) is the foundational virtual and physical platform that enables CSPs to deploy virtual network functions (VNFs) or cloud native functions (CNFs) across their network, including network services, operations support systems, business support systems and many types of IT and IoT applications.

Ericsson NFVi pre-integrates the latest cloud and open-source technologies, including software-defined infrastructure and support for network slicing and automated orchestration into a standards-based telco-grade platform ready for 4G, 5G and beyond. Learn more about Ericsson NFVI.

The emergence of NFV and cloud native has powered a new range of technological possibilities on the network, forming one of the key building blocks to many advanced 5G use cases today.

5G Core

5G core is responsible for managing data, signaling, and connections between users and services. Based primarily on cloud-native technology, 5G core supports advanced features like network slicing, edge computing, and ultra-low latency to enable diverse use cases from enhanced mobile broadband (eMBB) to massive IoT and mission-critical communications.

Network slicing

5G’s virtualized and cloud-native architecture is one of the underlying enablers that makes it possible to run multiple virtual networks called slices on a common physical infrastructure. These slices can be logically separated with different performance characteristics, enabling new differentiated connectivity models and exciting innovative possibilities.

Edge computing

Increasing demand from industries and enterprises for new use cases are driving demand for a new kind of programmable infrastructure with an increased focus on data-centric processing, security, response time, scalability and resilience. NFV and cloud-native architecture makes this possible, enabling CSPs to distribute network functions closer to the edge for near-real-time processing.

Network and service exposure

NFV makes it possible for CSPs to expose network functions through open APIs, boosting the programmability and adaptability of connectivity services to fit different needs and enabling new value creation across the whole ecosystem.

The business case: Launching Australia’s first virtualized core network on a multi-vendor capable cloud infrastructure with TPG Telecom

As a market challenger, TPG Telecom recognized the vital role that cloud-native and 5G Standalone would be play in opening exciting new market possibilities for Australian subscribers and enterprises in coming years.

Together with Ericsson, TPG Telecom took a historic step closer to making those possibilities real with the launch of the market’s first commercial 5G standalone (SA) network – complete with Ericsson’s multi-vendor cloud-native NFVI architecture.

Learn more about TPG Telecom cloud transformation journey here.

Video: Giovanni Chiarelli, Chief Technology Officer at TPG Telecom, shares highlights from their cloud transformation journey together with Ericsson.

NFV management, automation and orchestration (MANO) is the architectural framework that manages and orchestrates the allocation of resources demanded by virtual network functions (VNFs). This includes lifecycle management and orchestration of the virtualized resources for the VNF, but also traditional aspects such as fault-, configuration-, accounting-, performance- and security management (FCAPS) of the VNF itself.

MANO ensures that dynamic infrastructural workflows are maintained across the network, meaning that optimal performance, flexibility and scalability can be delivered at all times.

It is comprised of three key building blocks: the orchestrator, the network function manager and the infrastructural manager.

NFV Orchestrator (NFVO)

NFV Orchestration enables automation of a hybrid network of physical, virtual and cloud native infrastructures including VNFs, supporting both resource orchestration and VNF lifecycle management orchestration.

VNF Manager (VNFM)

The VNF manager (VNFM) performs seamless lifecycle management operations like instantiation, scale-in/scale-out, termination, upgrade and healing, together with the NFVO layer. The VNFM acts as a layer between service orchestration and the container management layer. This forms a modular and layered orchestration architecture prepared for higher layer service orchestration.

Virtualized Infrastructure Manager (VIM)

The virtual infrastructure manager (VIM) dynamically allocates resources for VNFs with telecom grade capabilities such as redundancy, high availability, high throughput with low latency, trusted tenant isolation, automatic virtual machine recovery and more.

Continuous integration and delivery (CI-CD)

As software, VNFs and the hosting commodity infrastructure require continuous integration, continuous delivery and continuous deployment (CI/CD) to ensure that every piece of software – from infrastructure layer, to VNF/CNF layer, to service layer – is developed, validated and deployed on a continuous basis, ensuring the compatibility and interoperability across the vertical layers of the stack.

Above: Lifecycle management of virtual- and cloud-native network functions.

Manage 5G workloads the hybrid way

Running 5G Core on any infrastructure requires a hybrid approach. Hybrid cloud management brings service provider- and hyperscale cloud infrastructure under one umbrella – delivering efficient and secure cloud operations with agility and flexibility.

Learn more

Persons discussing how to work with hbrid cloud

Cloud native is the next architectural revolution that is transforming how network functions are deployed and managed. This shift is happening now, with many of the world’s CSPs having already launched cloud native platforms in their core networks today.

Kubernetes over virtualized infrastructure vs. bare metal infrastructure

NFV: a stepping- stone to the cloud native revolution

In most cases, the shift from NFV to cloud native is a gradual one. By deploying a container-as-a-service (CaaS) platform, such as Kubernetes, on top of the exiting virtual infrastructure manager (VIM), cloud-native functions can co-exist with both legacy and virtual network functions.

NFV vs cloud native: the key differences

While NFV virtualizes network services traditionally run on dedicated hardware, cloud-native infrastructure is designed to run applications natively in the cloud using microservices, containers and other cloud-native technologies.

	NFV	Cloud-native
Architecture	Virtual Network Functions (VNFs) running on virtual machines (VMs) through a hypervisor (virtualization layer) on data center (DC) hardware typically standardized servers (COTS).	Microservices deployed within containers, orchestrated by platforms such as Kubernetes. They can run on bare metal infrastructure or on top of VMs but bare metal is the most efficient.
Scalability	Scalable VNFs by adding more VMs	Fine-grained scalability through individual microservices and containers
Resource efficiency and deployment speed	VM hypervisor layer results in a higher overhead and slower setup and instantiation times	Lightweight containers running directly on bare metal infrastructure removes the need for a hypervisor layer, reducing overhead and improving deployment speed
Resilience and fault tolerance	Resilience through VM failover mechanisms	Designed for resilience with self-healing, container orchestration and microservices redundancy
Operational model	Typically managed using traditional IT and network management tools	Utilizes DevOps practices, CI/CD pipelines and cloud-native management tools.
Adaptability and flexibility	Generally less flexible; changes often require redeployment of VMs.	Highly adaptable; microservices can be updated independently without impacting the entire system.
Development cycle	Monolithic nature of VNFs results in a slower innovation cycle	Automated CI/CD pipelines and microservice architecture contribute to faster innovation cycles
Cost efficiency	Potentially higher due to VM resource needs and overheads	Typically more cost-efficient with better resource utilization and dynamic scaling

Cloud native infrastructure based on bare metal is the future

The introduction of cloud-native bare metal infrastructure removes the need for network functions to run on virtual machines over a virtualization layer. Hosting containers directly on cloud-native bare metal infrastructure, as opposed to virtualized infrastructure, eliminates the need for additional infrastructure overheads such as the host operating system, hypervisor, guest operating system, SDN controller and more. This results in higher resource utilization, simplified operations and significant OPEX and CAPEX savings.

With network functions that are fully cloud native, decomposed and aligned with the Cloud Native Computing Foundation (CNCF) ecosystem, and a simplified architecture common for central, edge and private network deployments, CSPs will be perfectly primed to harness the possibilities of a cloud-native future.

Explore now

Ericsson cloud native infrastructure

Explore more insights about cloud native

Bare metal cloud: your questions answered

What is bare metal, and why is it so important for telecom? Find out in this insightful podcast with Ericsson's Head of NFVI Portfolio Strategy Management.

Why Kubernetes over bare metal is optimal for cloud native

Learn why Kubernetes is the best choice for deploying 5G cloud native applications and why some myths just aren't true.

Level up your network with our cloud native guide series

Explore the series