PaperlayerView exact revision on arXiv

Source-linked AI summary

NFV and SDN - Key Technology Enablers for 5G Networks

Faqir Zarrar Yousaf, Michael Bredel, Sibylle Schaller, Fabian Schneider

arXiv:1806.07316v1cs.NI

TL;DR

5G requires a flexible, scalable, agile, and programmable platform for services with diverse requirements, creating challenges for network architecture and management. The paper surveys NFV and SDN in relation to 5G, explaining their architectures and complementarity. It presents them as key technology enablers for realizing 5G networks and describes their use in a 5G core network.

  • Problem

    5G must provide diverse services within strict performance bounds while remaining flexible, scalable, agile, programmable, and cost efficient, requiring major architectural and technological redesign.

  • Method

    The paper provides an overview of NFV and SDN, describes how they can realize a 5G core network, and discusses how they complement each other.

  • Results

    NFV and SDN are presented as key technological enablers for realizing 5G networks, with NFV addressing virtualized-service flexibility and SDN enabling programmable connectivity.

  • Takeaways & Limitations

    The paper’s supported conclusion is that NFV and SDN together provide complementary capabilities for deploying and operating 5G network services.

Abstract

from arXiv · show

Communication networks are undergoing their next evolutionary step towards 5G. The 5G networks are envisioned to provide a flexible, scalable, agile and programmable network platform over which different services with varying requirements can be deployed and managed within strict performance bounds. In order to address these challenges a paradigm shift is taking place in the technologies that drive the networks, and thus their architecture. Innovative concepts and techniques are being developed to power the next generation mobile networks. At the heart of this development lie Network Function Virtualization and Software Defined Networking technologies, which are now recognized as being two of the key technology enablers for realizing 5G networks, and which have introduced a major change in the way network services are deployed and operated. For interested readers that are new to the field of SDN and NFV this paper provides an overview of both these technologies with reference to the 5G networks. Most importantly it describes how the two technologies complement each other and how they are expected to drive the networks of near future.

I. INTRODUCTION

5G must support diverse services, devices, and stringent performance requirements across a redesigned network architecture. The paper presents slicing, NFV, and SDN as responses to the resulting management, orchestration, scalability, and programmability challenges.

  • 5G Vision: 5G is expected to support enhanced mobile broadband, massive machine-type communication, and ultra-reliable low-latency communication.The requirements include Gbps-per-user peak rates, improved spectrum efficiency and coverage, and many diverse connected devices.
  • 5G Slicing: Network slicing enables multiple logical, self-contained networks to share one physical infrastructure while meeting different service requirements.Slices can be separately orchestrated and configured, with tenants controlling physical, virtualization, and service layers.
  • Challenges: The multi-tiered 5G architecture creates challenges for managing physical and virtual resources, orchestrating services, and providing programmable end-to-end connectivity.Slices must be deployed and managed across the resource, slice-instance, and service-instance layers.
  • Technology Enablers: NFV and SDN are identified as key technologies for 5G: NFV addresses flexible, agile, and scalable virtualized services, while SDN makes connectivity programmable.The paper examines their architectures, realization of a 5G core network, and complementarity.

II. NFV AND MANO SYSTEMS

NFV moves network functions from dedicated hardware into software and virtual machines, using cloud-oriented management to improve flexibility and automate operations. However, early appliance-style designs do not fit cloud performance, scaling, and availability assumptions, motivating smaller cloud-native components.

  • Network Function Virtualization: NFV virtualizes functions such as routers, firewalls, and evolved packet cores so they run as software rather than dedicated specialized hardware.This shift enables common hardware and cloud-management practices for network operations.
  • Network Function Virtualization: Using common off-the-shelf hardware and cloud-management tools can reduce capital and operational expenditures through automation.NFV also aims to incorporate failure management and horizontal partitioning into the system architecture.
  • Network Function Virtualization: Early NFV implementations packaged each network function as a large virtual machine and chained multiple VNFs into network services.Service Function Chains determine how packets are forwarded between VNFs.
  • Challenges: Fat virtual machines and traditional availability mechanisms often perform poorly in cloud environments because of I/O overhead, limited horizontal scaling, and mismatched failure assumptions.Cloud-native systems may kill and restart service instances at any time, unlike legacy availability designs.
  • Challenges: Current approaches favor smaller container-based network applications, but this substantially increases the number of entities that NFV management must orchestrate.The shift toward cloud-native design therefore intensifies management and orchestration challenges.

B. Management and Orchestration

MANO coordinates virtualized infrastructure, network connectivity, VNFs, and their life cycles across the multi-tiered 5G architecture. The ETSI framework organizes these responsibilities among infrastructure management, VNF management, and network-service orchestration functions.

  • Management and Orchestration: MANO systems simplify complex NFV services by coordinating cloud, network, VNF, and component life-cycle management.In 5G, resource, network-slice-instance, and service-instance layers must be managed together.
  • ETSI MANO Framework: The ETSI MANO reference model contains three main functional blocks: VIM, VNFM, and NFVO.They correspond to infrastructure management, VNF management, and NFV orchestration layers.
  • ETSI MANO Framework: The NFVO onboards network-service descriptions and manages service life cycles while maintaining end-to-end integrity across interconnected VNFs.It can connect to external OSS and BSS systems.
  • ETSI MANO Framework: The VNFM manages individual VNF life cycles, including actions such as starting, scaling, and configuring VNFs.It may interact directly with element managers and VNFs.
  • ETSI MANO Framework: The VIM manages virtual machines and containers and provides connectivity between VNFs by establishing virtual links, potentially through SDN.ETSI descriptors such as NSDs and VNFDs provide life-cycle and monitoring metadata for executing services and functions.

C. MANO Implementations

Open-source and commercial projects implement or extend the ETSI MANO framework, although many remain at an early stage. OSM is presented as an operator-driven ETSI reference implementation built from components covering the MANO layers.

  • MANO Implementations: Several open-source and commercial projects target MANO implementations, but most are still in an early stage.These projects nevertheless demonstrate holistic service management and orchestration for NFV.
  • OSM: OSM is an ETSI project intended as an operator-driven reference implementation of the ETSI MANO specification.Its components include a VIM connector, Canonical JuJu, and Rift.ware.
  • OSM: OSM maps its components onto virtual infrastructure management, VNF management, and NFV orchestration layers.The VIM connector supports multiple VIMs, JuJu manages life cycles, and Rift.ware performs end-to-end service delivery.

ONAP - Open Network Automation Platform:

The paper reviews several open-source MANO platforms for deploying, orchestrating, and managing virtualized network services in 5G networks. These platforms differ in architectural scope, lifecycle support, development tooling, and infrastructure integration.

  • ONAP - Open Network Automation Platform:: ONAP aims to provide a comprehensive framework for real-time, policy-driven automation of virtual network functions.It evolved from Open-O and ECOMP and was governed by the Linux Foundation, although its code was still under heavy development.
  • OpenStack Tacker:: OpenStack Tacker provides an ETSI MANO-based orchestrator with VNF deployment, lifecycle, configuration, and health-monitoring capabilities.Its feature set includes a VNF catalog and a general-purpose VNF Manager for end-to-end operation on an NFV platform.
  • OpenBaton:: OpenBaton implements ETSI MANO through an NFV Orchestrator, a generic VNF Manager, and an SDK for building specialized VNF Managers.Its orchestrator uses OpenStack as a VIM and can deploy multiple slices in parallel across cloud instances, supporting a multi-tenant environment.
  • SONATA - Agile Service Development and Orchestration in 5G Virtualized Networks:: SONATA combines service development tools, a service development kit, and orchestration in a framework for virtualized services.Its MANO emulator enables testing complex services on one computer, while its modular microservice architecture allows platform components such as plugins to be replaced.

D. Management and Orchestration of 5G Slices

5G slice management requires coordinated orchestration across infrastructure, slice, and service layers. NFV MANO supplies corresponding building blocks and extends management beyond traditional FCAPS to lifecycle, policy, performance, and resource functions.

  • D. Management and Orchestration of 5G Slices: ETSI NFV MANO maps the VIM, VNFM, and NFVO to the infrastructure, network slice manager, and service instance layers of 5G slicing.This correspondence indicates that ETSI MANO contains the building blocks needed for a 5G network-slice MANO framework.
  • D. Management and Orchestration of 5G Slices: A 5G MANO framework must coordinate multiple complex tasks to provision network slices beyond traditional FCAPS management.Additional functions include software-image, reliability, policy, bandwidth and latency, QoS/QoE, mobility, energy, charging, and infrastructure management.
  • D. Management and Orchestration of 5G Slices: Network-slice lifecycle management operates by managing the individual VNFs that compose each slice.The VNF is identified as the basic building block of a network slice at the virtualization layer.

III. SOFTWARE DEFINED NETWORKING

SDN addresses the need for agile traffic engineering and continuous network reconfiguration in NFV-enabled 5G networks. It does so by separating control from forwarding, exposing programmable abstractions, and using controllers to mediate between applications and resources.

  • III. SOFTWARE DEFINED NETWORKING: NFV introduces fast, agile connection setup and continuous equipment updates that are difficult to achieve with traditional network operations.SDN is presented as a way to overcome these limitations for connecting VNFs and establishing end-to-end services.
  • III. SOFTWARE DEFINED NETWORKING: SDN replaces rigid proprietary network equipment and services with programmable, software-driven functions spanning multiple vendor platforms.This decouples agile software release cycles from the slower release cycles of integrated hardware and software.
  • III. SOFTWARE DEFINED NETWORKING: SDN abstractions let operators manage services through lower-level functionality, supporting rapid service delivery and more efficient resource use.Deploying agile software on white-box switches is expected to improve network cost-performance behavior.
  • A. OpenFlow and ONF: OpenFlow enables communication between infrastructure elements and software-based network controllers.The protocol is maintained by the ONF and supported by major network equipment vendors.
  • III. SOFTWARE DEFINED NETWORKING: The ONF SDN architecture specifies decoupled control and forwarding, logically centralized control, and programmable network services.Applications use defined APIs to act on exposed resource abstractions and state.
  • III. SOFTWARE DEFINED NETWORKING: SDN controllers provision, manage, and control services by mediating between applications through northbound interfaces and resources through southbound interfaces.Authorized applications can invoke services or change network behavior through management-control sessions.

B. SDN implementation

SDN implementation spans programmable switching hardware, white-box devices, operating systems, and software controllers. The architecture places the controller between control applications and network resources.

  • B. SDN implementation: SDN system development includes a growing range of commercial and non-commercial switches, routers, and SDN controllers.The paper presents these as examples from an expanding implementation ecosystem.
  • B. SDN implementation: Early SDN switches translated protocol abstractions into switching-chip tables or used programmable hardware such as NPUs and FPGAs.Newer chips increasingly support programmable forwarding planes natively.
  • B. SDN implementation: The ONF architecture places the SDN controller centrally between control applications and network resources.Northbound interfaces connect applications to the controller, while southbound interfaces connect it to resources.
  • B. SDN implementation: Bare-metal or white-box switches permit installation of different operating systems instead of shipping with proprietary software.The operating-system market for these switches is evolving rapidly with a wide range of available options.

SDN Software:

SDN controllers are available in both commercial and open-source forms, with OpenDaylight and ONOS highlighted as prominent examples.

  • OpenDaylight and ONOS are prominent SDN controller examples governed by the Linux Foundation.

C. Future directions in SDN research

SDN research is extending programmability beyond conventional forwarding rules, especially to support new protocols and NFV deployments. These extensions can move part of VNF functionality into network elements, reducing virtual-machine requirements.

  • SDN research seeks to expand forwarding beyond existing data-plane protocols and legacy network-operation approaches.Contemporary implementations primarily support established header fields and typical actions such as forwarding or dropping packets.
  • Flexible matching on arbitrary header fields is being pursued through PIF/P4, POF, and DPN.These approaches address limitations users face when defining new data-plane protocols.
  • Action programmability is also being extended, including in-switch packet generation triggered by packets or other in-switch events.
  • These SDN extensions can partially implement VNF functionality directly in network elements, reducing the need to deploy virtual machines.

IV. THE MARRIAGE OF SDN AND NFV

The paper presents SDN and NFV as complementary technologies whose integration requires reconciling cloud/IT and telecom control models. It describes unified infrastructure control, holistic service descriptions, and multiple ETSI NFV integration options.

  • The marriage of SDN and NFV: SDN–NFV integration must reconcile the cloud, data-center, and IT view with the telecom operator and vendor view.The paper identifies these distinct communities as producing different chains of command.
  • The marriage of SDN and NFV: A unified infrastructure resource controller can combine virtual-machine management and network-controller functions on the same level.The paper notes that this model is implemented in SONATA and supported by the ONF SDN architecture when compute and storage are included.
  • The marriage of SDN and NFV: Holistic service descriptions should specify application components, deployment topology, scaling hints, network-path requirements, network functions, and programmable interfaces.
  • The marriage of SDN and NFV: ETSI NFV defines a precise framework for managing and orchestrating NFV infrastructure resources used by telecom functions and services.ONF takes a broader view of network systems, allowing SDN to map to varied specifications and implementations.
  • The marriage of SDN and NFV: ETSI NFV is specifying orchestration interfaces between the SDN controller and NFV MANO while accounting for interfaces internal to the SDN domain.
  • The marriage of SDN and NFV: ETSI NFV describes five options for positioning the SDN controller, including OSS/BSS, NFV Infrastructure, a physical network function, or a VNF.The cited passage introduces five options; the supplied text lists four positions explicitly.
  • The marriage of SDN and NFV: NFV can use SDN programmability to implement forwarding graphs, policy-based VNF chains, and traffic-path changes based on VNF processing.
  • The marriage of SDN and NFV: In a multi-site scenario, the NFV Orchestrator requests connectivity from VIMs and a WIM, which can request virtual networks from an SDN network controller.The scenario includes two NFVI-PoPs, each with its own VIM, plus WAN connectivity managed through a WIM.

V. CONCLUSION

The conclusion frames NFV and SDN as key enablers for managing complex, programmable 5G networks, while emphasizing that major research and interoperability challenges remain open.

  • 5G networks must flexibly serve massive numbers of users and devices and support new applications and services in an agile, programmable manner.
  • NFV and SDN are widely considered key enablers for 5G network architecture, design, operation, and management.Standardization organizations are developing architecture frameworks and interfaces for combining system components.
  • Open-source projects such as OpenStack, OPNFV, OSM, ONAP, ODL, and ONOS are expected to complement or replace commercial products.The projects and standardization bodies mutually influence one another while developing and validating their work.
  • Open research challenges include managing software agility, distributing functions across execution platforms, and ensuring interoperability among vendors.
  • Further challenges include mapping tenant business requirements to provider infrastructure and fulfilling QoS, QoE, and SLA requirements in cloud-native environments.
  • 5G slicing raises resource-management and automation questions across slices that may share or cross administrative domains.
Loading 1806.07316v1…