Source-linked AI summary
Network Slicing for 5G with SDN/NFV: Concepts, Architectures and Challenges
Jose Ordonez-Lucena, Pablo Ameigeiras, Diego Lopez, Juan J. Ramos-Munoz, Javier Lorca, Jesus Folgueira
TL;DR
5G needs to support vertical industries whose use cases have diverging requirements that current one-size-fits-all networks cannot address. The paper studies network slicing with SDN and NFV, analyzes their architectural roles, and illustrates an integrated deployment. It concludes that combining these technologies can realize slices, while performance, orchestration, security, privacy, and transition issues remain open challenges.
Problem
Vertical industries generate diverse use cases whose latency, scalability, availability, and reliability requirements exceed the flexibility of one-size-fits-all networks.
Method
The paper studies SDN and NFV architectures, analyzes their integration for slicing, and presents an SDN-enabled NFV deployment example.
Results
The paper presents a deployment in which SDN and NFV functional blocks, controllers, and interactions realize network slices on shared infrastructure.
Takeaways & Limitations
Network slicing provides a framework for accommodating diverse business-driven use cases through self-contained logical networks on common infrastructure.
Takeaways & Limitations
Shared infrastructure makes performance isolation difficult, while security and privacy concerns remain a major barrier to multi-tenancy adoption.
Abstract
from arXiv · showhide
The fifth generation of mobile communications is anticipated to open up innovation opportunities for new industries such as vertical markets. However, these verticals originate myriad use cases with diverging requirements that future 5G networks have to efficiently support. Network slicing may be a natural solution to simultaneously accommodate over a common network infrastructure the wide range of services that vertical-specific use cases will demand. In this article, we present the network slicing concept, with a particular focus on its application to 5G systems. We start by summarizing the key aspects that enable the realization of so-called network slices. Then, we give a brief overview on the SDN architecture proposed by the ONF and show that it provides tools to support slicing. We argue that although such architecture paves the way for network slicing implementation, it lacks some essential capabilities that can be supplied by NFV. Hence, we analyze a proposal from the ETSI to incorporate the capabilities of SDN into the NFV architecture. Additionally, we present an example scenario that combines SDN and NFV technologies to address the realization of network slices. Finally, we summarize the open research issues with the purpose of motivating new advances in this field.
1. Introduction
5G must support vertical industries whose use cases impose divergent requirements that one-size-fits-all networks cannot efficiently satisfy. The paper presents network slicing, enabled by SDN and NFV, as an approach for deploying tailored logical networks over shared infrastructure.
- Vertical industries such as manufacturing, automotive, healthcare, energy, and media impose requirements that differ substantially from existing services.
- One-size-fits-all networks cannot address diverging latency, scalability, availability, and reliability requirements across vertical use cases.
- Network softwarization uses SDN and NFV to create multiple logical networks tailored to different use cases on common infrastructure.
- A network slice is an on-demand, mutually isolated E2E logical network with independent control and management over a shared physical or virtual network.
- The paper studies SDN and NFV architectures, presents an integrated deployment example, and identifies challenges for 5G network slicing.
2. Background on key concepts for Network Slicing
Network slicing combines network functions and infrastructure resources through virtualization, while orchestration coordinates their use across diverse actors. Strong isolation is required so slices can meet service-specific performance, security, and privacy requirements on shared substrates.
- 2.1 Resources: A slice combines network functions and infrastructure resources whose capabilities are selected and composed to meet a use case’s service requirements.
- 2.2 Virtualization: Virtualization abstracts and logically partitions heterogeneous resources, enabling them to be shared among slices as virtual resources.
- 2.2 Virtualization: Virtualization introduces infrastructure providers, tenants, and end users with distinct roles in delivering network services.
- 2.2 Virtualization: Virtualization can be recursive: a tenant at one layer may act as an infrastructure provider for another tenant at the layer above.
- 2.3 Orchestration: Orchestration continuously selects and coordinates resources to fulfill client demands under optimization policies, changing demands, and available resources.
- 2.4 Isolation: Strong isolation requires independent performance, security, and privacy across slices, enforced through consistent policies and mechanisms alongside virtualization and orchestration.
3. ONF Network Slicing Architecture
The ONF SDN architecture supports slicing by dynamically abstracting and configuring forwarding resources for tailored client services. Its controller uses virtualization, orchestration, contexts, and recursion to construct and manage slices across abstraction layers.
- ONF SDN architecture: The ONF architecture places a logically intermediate control plane between forwarding resources and applications, dynamically configuring abstracted resources for tailored services.This alignment makes SDN suitable for agile, cost-effective support of diverse 5G slicing demands.
- Controllers and contexts: Controllers mediate between clients and resources, operating infrastructure resources, network functions, and network services through logically centralized control.The architecture applies the recursion principle by treating services as resources at higher abstraction levels.
- Virtualization and orchestration: SDN controllers virtualize and orchestrate resources into client-specific Resource Groups that provide each client with an abstract, customized service view.Virtualization abstracts and partitions resources, while orchestration dispatches selected resources to separate groups.
- Administration: Administrators instantiate and configure controllers, including their server and client contexts and associated policies.These contexts organize the resources and control logic exposed to clients.
- Slicing and recursion: A client context naturally represents a network slice by combining an abstract Resource Group, control logic, and related client service attributes.Recursion also enables multiple hierarchically arranged controllers and recursive composition of slices.
4. NFV Reference Architectural framework
The paper presents NFV as complementing SDN by managing slice-resource lifecycles and VNF allocation. An ETSI-inspired integration places SDN controllers at tenant and infrastructure levels, providing distinct overlay and underlay resource views.
- SDN–NFV complementarity: SDN supplies control-plane functionality for slicing but lacks capabilities needed to efficiently manage network-slice lifecycles and constituent resources.NFV addresses this gap by managing infrastructure resources and orchestrating those required for VNFs and network services.
- Integrated framework: The proposed common framework integrates SDN into NFV through two controllers, one logically located in the tenant domain and one at the infrastructure-provider level.The integration is intended to combine SDN control with NFV management and orchestration.
- NFV entities: NFV comprises NFVI for hosting and connecting VNFs, VNFs as software implementations of network functions, and MANO for virtualization-specific management and automation.MANO includes the VIM, VNFM, and orchestrator functions.
- Controller roles: The infrastructure SDN controller manages networking resources for VNF connectivity under VIM direction, while the tenant SDN controller manages tenant-domain resources.Both controllers operate through programmable southbound interfaces, including OpenFlow, NETCONF, or I2RS.
- Resource views: The infrastructure controller exposes an underlay for VNF deployment and connectivity, whereas the tenant controller exposes an overlay of composed VNFs defining tenant-managed network services.Their differing abstraction levels shape how each controller operates and what it presents through its interfaces.
5. Network Slicing use case with SDN-NFV Integration
The deployment example combines SDN and NFV to run multiple tenant-managed slices over common NFVI. Per-slice management components provide isolation, while orchestration allocates shared resources across divergent slice requirements at infrastructure and tenant levels.
- Deployment scenario: The example deploys several slices on common NFVI, with two tenants independently managing slices composed of chained VNFs that realize network services.It assumes a single recursion level, so tenants directly serve end users.
- Infrastructure deployment: The scenario uses compute and networking resources from one infrastructure provider across two NFVI-PoPs, plus SDN-based WAN transport from two additional providers.Infrastructure controllers programmatically manage connectivity under VIM and WAN infrastructure-manager direction.
- Slice management: Each slice contains its own OSS, tenant controller, and network-service orchestrator, enabling users to operate services within tenant-defined limits.The separate NSO, OSS, and TC instances provide the required management isolation.
- Resource orchestration: The resource orchestrator allocates tenant resources to slices while sharing them according to tenant and slice policies and required performance levels.Its framework is intended to enable performance isolation despite shared-resource operation.
- Two-level management: Resource management occurs at infrastructure and tenant levels, with VIM/WIM components supplying virtualized infrastructure and the RO assigning resources to slices.Both levels collect usage information and forecast near-term resource availability to satisfy demands.
- Modularity and security: The deployment models functional blocks as independent software components connected through APIs, while applying compartmentalization and per-resource security mechanisms across virtualization levels.Recursion can be added by allowing some slice users to act as tenants operating their own slices.
6. Challenges and Research directions
The paper identifies challenges in enforcing performance isolation, orchestrating dynamic multi-tenant resources, securing open interfaces, and adapting business and regulatory models for 5G slicing.
- Performance issues: Shared infrastructure makes performance isolation difficult: dedicated resources avoid interference but cause over-provisioning, whereas sharing can leave slices coupled in performance.Resource-management mechanisms must share resources when necessary without violating required performance levels.
- Management and orchestration: Dynamic, scalable slicing makes multi-tenant management and orchestration difficult because resource demands can vary considerably over short timescales.Proposed directions include cooperation between slice-specific blocks and the resource orchestrator, automatically validated policies, and efficient allocation and conflict resolution.
- Security and privacy: Open programmable interfaces create new potential attacks, requiring multi-level mechanisms for integrity, attestation, threat mitigation, authentication, and accounting.Security and privacy concerns are identified as a major barrier to adopting multi-tenancy approaches.
- New business models: New partnerships and tenants create promising business models, but transition strategies, regulatory review, pricing, cost sharing, and interoperability remain to be studied.The transition must support gradual evolution toward 5G while preserving compatibility with past infrastructure investments.
Biographies
The biographies describe three Telefónica professionals whose roles and research span network virtualization, 5G radio access, transport and IP networks, SDN/NFV, and related technologies.
- Diego Lopez: Diego Lopez leads Technology Exploration activities within Telefónica’s GCTO Unit and focuses on network virtualization, infrastructure services, network management, architectures, and security.He chairs the ETSI ISG on Network Function Virtualization and the NFVRG within the IRTF.
- Javier Lorca: Javier Lorca leads Radio Access Networks Innovation within Telefónica Global CTO and researches 5G technologies including virtualization, massive MIMO, mmWaves, waveforms, interference control, and channel coding.He holds an M.Sc. in Telecommunication Engineering and is pursuing a PhD at Universidad de Granada.
- Jesús Folgueira: Jesús Folgueira heads Transport and IP Networks within Telefónica’s Global CTO unit, overseeing network planning and technology while focusing on optical, metro, IP, SDN/NFV, and advanced switching.His education includes telecommunications engineering and telecommunication economics.