Vai al contenuto principale

Telecom Cloud and Network Infrastructure: Edge, ORAN, Slicing and Automation

Telecom Cloud and Network Infrastructure: Edge, ORAN, Slicing and Automation

This guide to telecom cloud and network infrastructure explains how modern networks are being rebuilt as software — cloud-native cores, edge computing, Open RAN, network slicing, and automation — and why that shift defines the 5G era. Telecom cloud and network infrastructure is the set of technologies that move network functions off dedicated hardware and onto containerised, orchestrated software running across central data centres and distributed edge sites, making the network programmable, sliceable, and automatable. This guide covers MEC, the cloud-native 5G core, Open RAN, slicing, and automation for operators, vendors, and enterprise architects.

TL;DR — key takeaways

  • Cloud-native networks run functions as containerised microservices on commodity/cloud infrastructure, orchestrated (e.g. Kubernetes) for agility and elasticity.
  • MEC (edge computing) puts compute near the user to cut latency — enabling automation, AR/VR, and real-time analytics.
  • Open RAN (ORAN) uses open interfaces so multi-vendor radio components interoperate, adding a RAN Intelligent Controller for optimisation.
  • Network slicing depends on cloud-native + edge to deliver isolated, SLA-backed virtual networks.
  • Automation (zero-touch, closed-loop, AI-driven) is essential as networks grow more distributed and complex.

What is telecom cloud and network infrastructure?

Telecom cloud and network infrastructure refers to the architecture that moves network functions off dedicated, proprietary hardware and onto software running across central data centres and distributed edge locations. Where a traditional network ran fixed appliances for each function, the modern network runs containerised, microservices-based software that can be orchestrated, scaled elastically, placed flexibly (centrally or at the edge), and updated continuously. This transformation spans the whole network: a cloud-native 5G core, virtualised and increasingly Open RAN radio, edge-computing platforms, and the automation and orchestration layers that tie them together. The shift makes the network programmable rather than static — the precondition for 5G’s defining capabilities of network slicing, low-latency edge services, and automated operations. It is as much a software-engineering and cloud-computing evolution as a telecom one, which is why hyperscalers, software vendors, and traditional network-equipment makers now all compete in this space.

What is multi-access edge computing (MEC)?

Multi-access edge computing (MEC) places compute and storage at the edge of the network — at a cell site, aggregation point, or local data centre — instead of in a distant centralised cloud. Processing data close to where it is generated and consumed slashes round-trip latency and reduces backhaul traffic, enabling applications that demand real-time response: industrial automation and robotics, autonomous vehicles, augmented and virtual reality, cloud gaming, and local video analytics. In 5G, MEC works hand in glove with the cloud-native core and network slicing to deliver low-latency, high-bandwidth services exactly where they are needed. For operators, edge computing is both a performance enabler and a new revenue line — offering enterprises local compute bundled with connectivity, often in partnership with hyperscale cloud providers. MEC is one of the principal reasons 5G is framed as a platform for new applications rather than merely a faster pipe, and it is central to industrial and enterprise 5G strategies.

The cloud-native 5G core: VNFs vs CNFs

The mobile core has evolved through three stages: physical appliances, then virtualised network functions (VNFs), and now cloud-native network functions (CNFs). VNFs — the network functions virtualisation (NFV) approach — run traditional network software on virtual machines on general-purpose servers, a real break from proprietary hardware but often still monolithic, ported rather than redesigned software. CNFs are designed from the ground up to be cloud-native: decomposed into microservices, packaged in containers, orchestrated by platforms such as Kubernetes, and able to scale and update individual components independently. A modern 5G standalone core is increasingly built as CNFs communicating over a service-based architecture, delivering far greater agility, elasticity, and resilience than VNF-based designs. The distinction matters because it determines how flexibly the network can scale to demand, heal itself, and roll out new features — the qualities that underpin slicing, edge services, and automated operations.

GenerationRuns onSoftware designScalingAgility
Physical appliancesDedicated hardwareMonolithic, proprietaryManual, hardware-boundLow
VNF (NFV)Virtual machinesOften ported monolithVM-levelMedium
CNF (cloud-native)Containers / KubernetesMicroservicesElastic, granularHigh

What is Open RAN (ORAN)?

Open RAN (ORAN) builds the radio access network from components connected by open, standardised interfaces, so hardware and software from different vendors can interoperate — breaking the traditional model in which a single vendor supplied the entire proprietary RAN as a tightly integrated stack. ORAN disaggregates the RAN into parts, notably the radio unit, distributed unit, and centralised unit, linked by open interfaces defined largely by the O-RAN Alliance, and adds a RAN Intelligent Controller (RIC) for software-driven optimisation and automation. Its goals are vendor diversity and reduced lock-in, more competition and innovation, the freedom to mix best-of-breed components, and software-based, virtualised, cloud-friendly radios that align with the broader cloud-native shift. ORAN is strategically important for supply-chain diversity — a significant policy concern in many markets — and for operator flexibility, though it adds integration and testing complexity that operators must weigh. It is one of the most actively debated topics in network infrastructure.

How network slicing depends on cloud and edge

Network slicing — running multiple isolated, purpose-built virtual networks over one physical 5G infrastructure — rests directly on cloud-native and edge foundations. Each slice is an end-to-end logical network spanning radio, transport, and core, and delivering one with guaranteed characteristics (latency, bandwidth, reliability) requires the underlying functions to be software-defined, virtualised, and dynamically orchestrated — exactly what a cloud-native core provides. Edge computing amplifies slicing’s value: a low-latency slice for factory automation or remote control is most effective when its supporting compute sits at the edge near the site rather than in a distant data centre. So slicing is the service-level capability, while cloud-native infrastructure and edge placement are the enabling foundations. Together they let operators sell differentiated, SLA-backed connectivity — effectively private-grade networks delivered as a service — which is a principal commercial rationale for investing in 5G standalone and the cloud-native infrastructure beneath it.

Why network automation is essential

Network automation uses software, orchestration, and increasingly AI/machine learning to design, deploy, configure, monitor, heal, and optimise networks with minimal manual effort. As networks become cloud-native, sliced, and distributed across central and edge sites, their complexity outstrips what manual operations can manage cost-effectively, making automation a necessity rather than a luxury. It spans zero-touch provisioning, closed-loop assurance (the network detects and corrects issues automatically), intent-based networking (operators specify desired outcomes and the system configures itself), and AI-driven optimisation of capacity, energy use, and performance. Automation controls operating cost as complexity rises, improves reliability and speed of service delivery, and is a prerequisite for delivering and assuring large numbers of slices and edge services at scale. It is a defining capability of modern telecom infrastructure, a major focus of operator investment, and a significant theme in 5G-Advanced standardisation, where AI is being woven directly into network management.

Related on DROAM

Frequently asked questions

What is a cloud-native telecom network?

A cloud-native telecom network is one whose functions — especially the mobile core — are built and run as containerised, microservices-based software on cloud infrastructure, rather than as monolithic applications on dedicated, purpose-built hardware. In a cloud-native 5G standalone core, network functions are decomposed into independent services that communicate over a service-based architecture, packaged in containers, orchestrated (commonly with Kubernetes), and able to scale elastically, update independently, and run on commodity servers or public/private cloud. The benefits are agility (new features and capacity deployed quickly), efficiency (resources scale to demand), resilience (self-healing and redundancy), and the ability to place functions flexibly — centrally or at the edge. Cloud-native is the architectural foundation that makes 5G’s programmability, network slicing, and automation possible; it is as much a software-engineering shift as a telecom one.

What is multi-access edge computing (MEC)?

Multi-access edge computing (MEC) places computing and storage resources at the edge of the network — close to where data is generated and consumed, such as at a cell site, aggregation point, or local data centre — rather than in a distant centralised cloud. By processing data near the user or device, MEC slashes round-trip latency and reduces backhaul traffic, enabling applications that need real-time response: industrial automation, autonomous vehicles, augmented and virtual reality, cloud gaming, and local video analytics. In the 5G context, MEC pairs with the cloud-native core and network slicing to deliver low-latency, high-bandwidth services where they are needed. For operators, edge computing is both a performance enabler and a new revenue opportunity — offering enterprises local compute alongside connectivity — and it is a key reason 5G is framed as a platform rather than just a faster network.

What is Open RAN (ORAN)?

Open RAN (ORAN) is an approach to building the radio access network using open, standardised interfaces between its components, so that hardware and software from different vendors can interoperate — breaking the traditional model where one vendor supplied the whole proprietary RAN as a tightly integrated stack. ORAN disaggregates the RAN into parts (notably the radio unit, distributed unit, and centralised unit) connected by open interfaces defined largely by the O-RAN Alliance, and adds a RAN Intelligent Controller (RIC) for software-driven optimisation and automation. The goals are vendor diversity and reduced lock-in, more competition and innovation, the ability to mix best-of-breed components, and software-based, virtualised, cloud-friendly radios. ORAN is strategically significant for supply-chain diversity and for operators seeking flexibility, though it adds integration complexity. It is a central theme in modern network infrastructure and policy discussions.

How does network slicing relate to cloud and edge?

Network slicing — running multiple isolated, purpose-built virtual networks over one physical 5G infrastructure — depends directly on cloud-native and edge technologies. Each slice is an end-to-end logical network spanning radio, transport, and core, and delivering a slice with guaranteed characteristics (latency, bandwidth, reliability) requires the underlying functions to be software-defined, virtualised, and dynamically orchestrated, which is exactly what a cloud-native core provides. Edge computing extends slicing’s value: a low-latency slice for, say, factory automation is most effective when its supporting compute sits at the edge near the factory rather than in a distant data centre. So slicing is the service-level capability, while cloud-native infrastructure and edge placement are the enabling foundations. Together they let operators offer differentiated, SLA-backed connectivity — effectively private-grade networks delivered as a service — which is one of the principal commercial rationales for 5G standalone.

What is network automation and why does it matter?

Network automation is the use of software, orchestration, and increasingly AI/machine learning to design, deploy, configure, monitor, heal, and optimise networks with minimal manual intervention. As networks become cloud-native, sliced, and distributed across central and edge sites, their complexity exceeds what manual operations can manage cost-effectively, so automation becomes essential rather than optional. It spans zero-touch provisioning, closed-loop assurance (the network detects and corrects issues automatically), intent-based networking (operators specify desired outcomes and the system configures itself), and AI-driven optimisation of capacity, energy, and performance. Automation matters because it controls operating cost as complexity grows, improves reliability and speed of service delivery, and is a prerequisite for delivering and assuring large numbers of network slices and edge services at scale. It is a defining capability of modern telecom infrastructure and a major focus of operator investment and 5G-Advanced standardisation.

What is the difference between virtualised and cloud-native network functions?

Virtualised network functions (VNFs) and cloud-native network functions (CNFs) are two generations of software-based networking. VNFs, the earlier approach (network functions virtualisation, NFV), take traditional network software and run it on virtual machines on general-purpose servers instead of dedicated appliances — a meaningful step away from proprietary hardware, but the software is often still monolithic and ported rather than redesigned. CNFs are designed from the ground up to be cloud-native: decomposed into microservices, packaged in containers, orchestrated by platforms like Kubernetes, and able to scale and update granular components independently. CNFs deliver greater agility, elasticity, and resilience than VNFs. The industry trajectory is from physical appliances, to VNFs on virtual machines, to CNFs in containers — with 5G standalone cores increasingly built as CNFs. Understanding the distinction matters because it determines how flexibly and efficiently a network can scale, heal, and evolve.