Individual S. Homma
Internet-Draft H. Nishihara
Intended status: Informational NTT
Expires: January 9, 2020 T. Miyasaka
KDDI Research
A. Galis
University College London
V. Ram OV
Independent Research Consultant India
D. Lopez
L. Contreras-Murillo
J. Ordonez-Lucena
Telefonica I+D
P. Martinez-Julia
NICT
L. Qiang
Huawei Technologies
R. Rokui
L. Ciavaglia
Nokia
X. de Foy
InterDigital Inc.
July 8, 2019
Network Slice Provision Models
draft-homma-slice-provision-models-01
Abstract
Network slicing is an approach to provide separate virtual network
based on service requirements. It's a fundamental concept of the 5G,
and the architecture and specification is under standardization in
several organizations. However, the definitions and scopes of
network slicing vary to some degree from one organization to another.
This document provides classification of provisioning models of
network slice for clarifying the differences on the definitions and
scopes.
Status of This Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at https://datatracker.ietf.org/drafts/current/.
Homma, et al. Expires January 9, 2020 [Page 1]
Internet-Draft draft-homma-slice-provision-models July 2019
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
material or to cite them other than as "work in progress."
This Internet-Draft will expire on January 9, 2020.
Copyright Notice
Copyright (c) 2019 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(https://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Table of Contents
1. Introduction and Motivation . . . . . . . . . . . . . . . . . 3
1.1. Differentiated Roles in Network Slice Provisioning . . . 3
1.2. High-level Problem Statement . . . . . . . . . . . . . . 4
2. Definition of Terms . . . . . . . . . . . . . . . . . . . . . 4
3. General Requirements for Network Slicing . . . . . . . . . . 7
3.1. Requirements/Attributes for Network Slice . . . . . . . . 7
4. Network Slice Structure . . . . . . . . . . . . . . . . . . . 8
4.1. Resources for Structuring Network Slices . . . . . . . . 8
4.2. Basic Network Slice Structure . . . . . . . . . . . . . . 12
4.3. Stakeholders in the Structuring Network Slices . . . . . 14
5. Variations of Network Slice Creation . . . . . . . . . . . . 15
5.1. Ready-made Network Slice . . . . . . . . . . . . . . . . 15
5.2. Custom-made Network Slice . . . . . . . . . . . . . . . . 15
5.3. semi-Custom-made Network Slice . . . . . . . . . . . . . 15
6. Network Slice Provision Models . . . . . . . . . . . . . . . 15
6.1. SaaS-like Model . . . . . . . . . . . . . . . . . . . . . 16
6.1.1. Capability in SaaS-like Model . . . . . . . . . . . . 16
6.2. PaaS-like Model . . . . . . . . . . . . . . . . . . . . . 16
6.2.1. Capability in PaaS-like Model . . . . . . . . . . . . 17
6.3. IaaS-like Model . . . . . . . . . . . . . . . . . . . . . 17
6.3.1. Capability in IaaS-like Model . . . . . . . . . . . . 17
6.4. Mapping of NS Provision Models and Infrastructure
Layering . . . . . . . . . . . . . . . . . . . . . . . . 17
7. Security Considerations . . . . . . . . . . . . . . . . . . . 19
Homma, et al. Expires January 9, 2020 [Page 2]
Internet-Draft draft-homma-slice-provision-models July 2019
8. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 19
9. Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . 19
10. Informative References . . . . . . . . . . . . . . . . . . . 19
Appendix A. NS Structure in the 3GPP 5GS . . . . . . . . . . . . 20
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 21
1. Introduction and Motivation
Network slicing is an approach to provide separate virtual networks
depending on requirements of each service. Network slicing receives
attention due to factors such as diversity of services and devices,
and it is also a fundamental concept of the 5G for applying networks
to such various types of requirements.
In addition, network slicing is expected to enable a business model
to provide dedicated logical networks to 3rd parties or vertical
customers on-demand, called NSaaS (Network Slice as a Service). For
such usage, in network slicing, provision of networks able to
guarantee communication characteristics end to end would be required.
However, the definitions are not harmonized over several SDOs
(Standards Developing Organizations).
This document clarifies provision patterns of network slice, and
provides the definitions and scope of network slicing which are
available over several organizations. Furthermore, the deliverables
would be help for evaluating applicabilities of existing
technologies/solutions to network slicing.
1.1. Differentiated Roles in Network Slice Provisioning
The widespread of system and network virtualization technologies has
conducted to new business opportunities, enlarging the offer of IT
resources in the form of Network Slices (NS). As a consequence,
there is a clear differentiation between the owner of physical
resources, the infrastructure operator, and the intermediary that
conforms and delivers network services to the final customers, the
Virtual Network Operator (VNO).
VNOs aim to exploit the virtualized infrastructures to deliver new
and improved services to their customers. However, current NS
techniques offer poor support for VNOs to control their resources.
It has been considered that the infrastructure operator is
responsible of the reliability of the NS elements but several
situations advocate the VNO to gain a finer control on its resources.
For instance, dynamic events, such as the identification of new
requirements or the detection of incidents within the virtual system,
might urge a VNO to quickly reform its virtual infrastructure and
resource allocation. However, the interfaces offered by current
Homma, et al. Expires January 9, 2020 [Page 3]
Internet-Draft draft-homma-slice-provision-models July 2019
virtualization platforms do not offer the necessary functions for
VNOs to perform the elastic adaptations they require to tackle with
their dynamic operation environments.
1.2. High-level Problem Statement
Beyond their heterogeneity, which can be resolved by software
adapters, NS platforms do not offer common methods and functions, so
it is difficult for the virtual network controllers used by the VNOs
to actually manage and control virtual resources instantiated on
different platforms, not even considering different infrastructure
operators. Therefore, it is necessary to reach a common definition
of the functions that should be offered by underlying platforms to
enable such overlay controllers with the possibility of allocate and
deallocate resources dynamically and get monitoring data about them.
Such common methods should be offered by all underlying controllers,
regardless of being network-oriented (e.g., ODL, ONOS, Ryu) or
computing-oriented (e.g., OpenStack, OpenNebula, Eucalyptus).
Furthermore, it is also important for those platforms to offer some
"PUSH" function to report resource state, avoiding the need for the
VNO's controller to "POLL" for such data. A starting point to get
proper notifications within current REST APIs could be to consider
the protocol proposed by the [WEBPUSH-WG].
Finally, in order to establish a proper order and allow the
coexistence and collaboration of different systems, a common ontology
regarding network and system virtualization should be defined and
agreed, so different and heterogeneous systems can understand each
other without requiring to rely on specific adaptation mechanisms
that might break with any update on any side of the relation.
2. Definition of Terms
This section lists definitions and terms related to network slicing.
This document refers terms and view points on network slicing in some
SDOs, such as 3GPP([TS.23.501-3GPP], [TS.28.530-3GPP], and
[TS.28.801-3GPP]), and NGMN ([NGMN-5G-White-Paper]). However the
scope of this document is not network slicing which is mobile
specific but one for general networks, and thus some of definitions
in this document may be different from ones of those documents.
Network Slicing: Network slicing indicates a technology, an
approach, or a concept to create logical separate networks in
support of services, depending on several requirements, on the
same physical resources. This is possible by combinations of
several network technologies.
Homma, et al. Expires January 9, 2020 [Page 4]
Internet-Draft draft-homma-slice-provision-models July 2019
Network Slice (NS): An NS is a logical separate network that
provides specific network capabilities and characteristics. In
3GPP definitions, an NS potentially includes both data plane and
control plane resources/functions.
Network Slice Instance (NSI): An NSI is a logical network instance
composed with required infrastructure resources, including
networking (WAN), computing (NFVI) resources, and some include
additional network service functions such as firewall or load-
balancer. It is composed of one or more Network Slice Subnet
Instances.
Network Slice Subnet: A Network Slice Subnet is a representation of
a set of required resources. It is composed and managed as a
group of network elements.
Network Slice Subnet Instance (NSSI): An NSSI is a partial logical
network instance represented as a network slicce instance. It is
a minimul unit managed or provided as a network slice. One or
more NSSI structure an NSI or an E2E-NSI.
End-to-End Network Slice Instance (E2E-NSI): An E2E-NSI is a virtual
network connecting among end points. It is composed of one or
multiple NSSIs. This term is original of this document and is
used when it should be emphasized that the target NSI provides
connectivity from end to end. As an example, for providing an
E2E-NSI on the 3GPP 5G network, combining three types of NSIs:
RAN-, TRN-, and CN-NSIs would be required.
Transport(TRN)-NSSI: A set of connections between various network
functions (VNF or PNF) with deterministic SLAs. They can be
implemented (aka realized) with various technologies (e.g. IP,
Optics, FN, Microwave) and various transport (e.g. RSVP, Segment
routing, ODU, OCH etc). The overview of NSI composed with TRN-
NSSI is shown in Appendix A.
RAN-NSSI: Regardless of RAN deployment (e.g. distributed-RAN,
Centralized-RAN or Cloud-RAN, a RAN-NSSI creates a dedicate and
logical resource on RAN for each NSI which are completely. The
overview of NSI composed with RAN-NSSI is shown in Appendix A.
Core(CN)-NSSI: Regardless of Core deployment, a CN-NSI creates a
dedicate and logical resource on Core network for each NSI which
are completely. The overview of NSI composed with CN-NSSI is
shown in Appendix A.
Network Slice as a Service (NSaaS): An NSaaS is a service delivery
model in which a third-party provider or a vertical customer hosts
Homma, et al. Expires January 9, 2020 [Page 5]
Internet-Draft draft-homma-slice-provision-models July 2019
NSIs and makes them available to customers. In this model, there
mainly two roles: NS provider and NS tenant.
Network Slice Provider (NS Provider): An NS provider is a person or
group that designs and instantiates one or more NSIs/NSSIs, and
provides them to NS tenants. In some cases, an NS provider is an
infrastructure operator simultaneously. This includes NSI, NSSI,
and E2E-NSI providers.
Network Slice Tenant (NS Tenant): An NS tenant is a person or group
that rents and occupies NSIs from NS providers.
Network Slice Stakeholder (NS Stakeholder): An NS stakeholder is an
actor in network slicing, and has roles of either NS provider or
tenant.
Infrastructure Operator: An infrastructure operator is an
organization who manages infrastructure networks or data centers
for running NSIs. In the most of cases, infrastructure operators
are initial NS providers on NSaaS. Also, some of them may be NS
tenants simultaneously.
Vertical Customer: A vertical customer is a organization who
provides some communicating services with using NSIs on NSaaS
model. In many cases, a vertical customer become the final NS
tenant on NSaaS. For example, video gaming companies or vehicle
vendors will possibly be vertical customers.
Virtual Network Operator (VNO): A VNO is a person or group that
operates virtual networks composed with resources or NSSIs rent
from infrastructure operators and provides such virtual networks
as NSIs to vertical customers who are final NS tenants. In some
cases, infrastructure operators have this role in addition to
operating their own infrastructure simultaneously.
Domain: A domain is a group of a network and devices administrated
under a policy-based common set of rules and procedures.
Resource: A resource is an element used to create virtual networks.
There are several types of resources, i.e., connectivity,
computing and storage. The details are described Section 4.1
Virtual Network: A virtual network is a network running a number of
virtual network functions.
Virtual Network Function (VNF): A virtual network function (VNF) is
a network function whose functional software is decoupled from
hardware. One or more VNFs run as different software and
Homma, et al. Expires January 9, 2020 [Page 6]
Internet-Draft draft-homma-slice-provision-models July 2019
processes on top of industry-standard high-volume servers,
switches and storage, or cloud computing infrastructure. They are
capable of implementing network functions traditionally
implemented via custom hardware appliances and middleboxes (e.g.,
router, NAT, firewall, load balancer, etc.).
Network Operation System: A network operation system is an entity or
a group of entities for operating network nodes and functions as
compositions of infrastructure network. For example, OSS/BSS,
orchestrator, and EMS are considered to be network operation
systems.
3. General Requirements for Network Slicing
On network slice operations, capabilities for dynamic instantiation,
change, and deletion should be required because an NSI is established
based on received orders from tenants in NSaaS. From this aspect,
some mechanisms to design a network based on service requirements and
to convert those to concrete configurations based on the design would
be required.
In addition, each NS has to maintain concrete communication
characteristics end to end, and resource reservations on data plane
and isolation among NSIs would be required. Isolation is a concept
to prevent the reduction of communication quality caused by
disturbance from other NSs, and it may have some levels of
enforcement, such as hard or soft isolations. In some cases, for
providing appropriate communication between client and server, it
would be allowed for NS tenants to put their applications as contents
server on NSIs by using computing resources.
The required agility of slice operation and granularity of end to end
communication quality requested can vary depending on provision
model.
3.1. Requirements/Attributes for Network Slice
NS tenants will have specific requirements for network slices
depending on the usages or service characteristics. Such
requirements or the assosiated attributes are broken down into
concrete design including network topology and configurations of
infrastructure resources, and NS is established based on the design.
The requirements or attributes on NSs are listed below:
o Requirements/Attributes of Network Resource
* bandwidth
Homma, et al. Expires January 9, 2020 [Page 7]
Internet-Draft draft-homma-slice-provision-models July 2019
* latency
* jitter
* packet loss rate
* reliability (e.g., MTBF, MTTF)
o Requirements/Attributes of Functionalities Resources
* function type (e.g., security, parental control)
* throughput
* packet error rate
* availability
4. Network Slice Structure
This section describes resources used for structuring NSs and the
basic structure of E2E-NS.
4.1. Resources for Structuring Network Slices
A network slice is structured as combinations of the resources it
uses. Such resources are mainly categorized into three classes:
network/WAN, computing/NFVI, and functionality resources. Variations
of each resources are described below. (Note that the lists are not
exhaustive.)
Network(WAN) Resources:
* Connectivity:
+ (v)Link
- Bandwidth per link/session
- Connected area/end points
- Forwarding route/path (e.g., for traffic engineering,
redundancy)
- Communication Priority (e.g., QoS class)
- Range of jitter amount
Homma, et al. Expires January 9, 2020 [Page 8]
Internet-Draft draft-homma-slice-provision-models July 2019
+ Interface of vNode
- QoS setting (e.g., Queue size, DSCP remarking, PIR/CIR)
- Filter setting
+ vRouter/vSwitch (# Treated as a set of (v)links and
interfaces of vNodes.)
* Multicast support
* Encryption support
* Authentication support
* Metadata conveyance (e.g., subscriber ID)
* Protocols for slice data plane:
+ VLAN
+ IPoE (IPv4 or IPv6)
+ MAP-E
+ DS-Lite
+ PPPoE
+ L2TP
+ GRE
+ MPLS
+ VxLAN
+ Geneve
+ GTP-U
+ Segment Routing MPLS
+ Segment Routing IPv6
+ NSH
+ Other
Homma, et al. Expires January 9, 2020 [Page 9]
Internet-Draft draft-homma-slice-provision-models July 2019
Computing(NFVI) Resources:
* (v)CPU core
* Storage
* Memory
* Disk
* vNIC
* Connectivity to VNF instances
* Virtual Deployment Unit:
+ Virtual Machine (VM)
+ container
+ micro kernel
* Resource Deployment Location (i.e., edge DC, central DC, public
cloud, ..., etc.)
Functionality Resources:
* Image:
+ Data Plane(DP) NF:
- GateWay(GW) function:
o Access Point Type (e.g., for radio, Wi-Fi, and fixed
accesses)
o Slice Selection Setting
o Terminate protocol
o Authentication
- Security Appliance:
o IPS (Intrusion Prevention System)
o IDS (Intrusion Detection System)
Homma, et al. Expires January 9, 2020 [Page 10]
Internet-Draft draft-homma-slice-provision-models July 2019
o WAF (Web Application Firewall)
- DPI
- Load Balancer
- TCP Accelerator
- Video Optimizer
- Parental Control
- Mobile DP functions (Ref. 3GPP 5GS)
gNB
UPF
Uplink Classifier
+ Control Plane(CP) NF:
- DHCP
o Fixed IP address allocation
o Dynamic IP address allocation
o The number of registered devices
- DNS
- VoIP (SBC, SIP server)
- Mobile CP function (Ref. 3GPP 5GS)
o AMF (Access and Mobility management Function)
o SMF (Session Management Function)
o PCF (Policy Control Function)
o UDM (Unified Data Management)
o NEF (Network Exposure Function)
* Provided VNF Type (e.g., open source, product of vender#A, ...,
etc.)
Homma, et al. Expires January 9, 2020 [Page 11]
Internet-Draft draft-homma-slice-provision-models July 2019
* Function location (e.g., edge DC, central DC, Public cloud,
etc.)
In terms of security or usability for NS tenants, some abstraction on
resource information would be required, however both setting
parameters of underlay infrastructure and abstracted information may
coexist in these lists.
For abstraction of parameters of underlay networks, some additional
protocols or functions (like [RFC8453] ) would be required.
Moreover, for providing strict communication qualities, combinations
of some technologies may be useful (ref.
[I-D.dong-teas-enhanced-vpn]).
4.2. Basic Network Slice Structure
An E2E-NSI is constructed by stitching NSSIs instantiated on each
participating domain. This includes the simplest case of a single
NSSI as an E2E NS. Domain types where some NSSIs are established are
described below:
o Fixed access network
o Mobile access network
o Transport network
o Fixed core network
o Mobile core network
o Data center (DC)
* Edge DC
* Central DC
o Private network
* Enterprise
* Factory
* Utilities
* Farming
* Home/SOHO
Homma, et al. Expires January 9, 2020 [Page 12]
Internet-Draft draft-homma-slice-provision-models July 2019
* Other
Figure 1 describes the overview of this structure. Resources in each
domain (e.g., access, core networks, and DC) are handled by
management entities and constitute an NSSI. An E2E-NSI is
established by stitching these NSSIs. Ways to stitch NS-subnets are
described in [I-D.defoy-coms-subnet-interconnection] and
[I-D.homma-nfvrg-slice-gateway].
___________________________________________
/ /
/__________________________________________/
E2E-NSI
A A A
| | |
____________ ___________ ______________
/ / / / / /
/___________/ /__________/ /_____________/
NSSI#1 NSSI#2 NSSI#3
A A A
| | |
o---o [PNF] /----[VM]
/ `--. / `----. /----[VM]
o---o-----o...o---------o...o---o----[VM]
NW Rsrc#1 NW Rsrc+PNF NW+CMP Rsrcs
A A A
| | |
,--------. ,--------. ,-----------.
/ Access \ / Core \ / \
| Network | | Network | | DC Domain |
\ Domain / \ Domain / \ /
`--------' `--------' `-----------'
*Legends
NW Rsrc : Network Resource
CMP Rsrc: Computing Resource
o : virtual/physical node structuring NSI
-- : virtual/physical link structuring NSI
[PNF]: Physical Network Function Appliance on NSI
[VM] : Virtual Machine Instance on NSI
Figure 1: Overview of NS Structure
Homma, et al. Expires January 9, 2020 [Page 13]
Internet-Draft draft-homma-slice-provision-models July 2019
Although it is shown that an NSSI belongs to just only one E2E-NSI in
Figure 1, it may be allowed that multiple E2E-NSIs share an NSSI.
Some resources may belong to multiple NSSI as well.
In addition, structure on composition of NSI may be recursive. In
other words, even though Figure 1 shows a case where NSSIs compose
directly an E2E-NSI, in some cases, NSSIs compose an NSI which is a
part of an E2E-NSI. The overview is shown in Figure 2. In this
figure, NSI#4 is composed of NSSI#1 and NSSI#2, and it structures
E2E-NSI#5 with NSSI#3.
___________________________________________
/ /
/__________________________________________/
E2E-NSI#5
A A
| |
___________________________ |
/ / |
/__________________________/ |
NSI#4 (= NSSI#1 + NSSI#2) |
A A |
| | |
____________ ___________ _____________
/ / / / / /
/___________/ /__________/ /____________/
NSSI#1 NSSI#2 NSSI#3
Figure 2: Overview of NS recursive structure
4.3. Stakeholders in the Structuring Network Slices
Potential stakeholders in network slicing are described below:
o NSSI provider: infrastructure operator
o Intermediate-NSI provider: infrastructure operator, VNO
o E2E-NSI provider: infrastructure operator, VNO, service provider
o NS tenant: infrastructure operator, VNO, service provider,
enterprise, mass user
o End customer: enterprise, mass user, etc.
Homma, et al. Expires January 9, 2020 [Page 14]
Internet-Draft draft-homma-slice-provision-models July 2019
5. Variations of Network Slice Creation
NSs can be classified according to their creation pattern into two
types: ready-made(RM) NS, custom-made(CM), and semi-custom-made(sCM)
NS. This section describes the features of these types.
5.1. Ready-made Network Slice
RM-NS is an NS creation pattern in which an infrastructure operator
decides service requirements by itself, and established based on the
requirements in advance. NS tenants select one of RM-NSs whose
features are closer to their requirements.
This model doesn't need immediacy on designing of NSI and enables to
mitigate the difficulty of implementation compared with other models.
5.2. Custom-made Network Slice
CM-NS is an NS creation pattern in which an NS is established based
on an order from a tenant and is provided to it. As examples of
usage of CM-NS, an enterprise builds and operates a virtual private
network for connecting several bases, or OTT (Over The Top) or other
industrial service providers create NSs based on their own
requirements and use them as a part of their own services (e.g.,
connected vehicles/drones, online video games, or remote surgery).
In this model, network operation system would be required to have
incorporate intelligence for designing appropriate NSs on-demand.
5.3. semi-Custom-made Network Slice
sCM-NS is a derivation of a CM-NS. In sCM-NS, an NS provider designs
the outline of NSs in advance, and a tenant tunes an NS with deciding
some parameters or applications run on resources. For example, an
infrastructure operator designs a logical network presenting
connectivity, and tenants install their own applications on servers
running on the logical network.
6. Network Slice Provision Models
This section classifis NS provision models into three categories
defined from aspect that granularity of information exposed to
tenants. The provision models are categorized into three models:
SaaS (Software as a Service) -like Model, PaaS (Platform as a
Service) -like Model, and IaaS (Infrastructure as a Service) -like
Model. The capabilities which NS tenants can have on management of
NSs would vary depending on the selected provision model.
Homma, et al. Expires January 9, 2020 [Page 15]
Internet-Draft draft-homma-slice-provision-models July 2019
6.1. SaaS-like Model
In SaaS-like Model, underlay infrastructure is hidden from tenants,
and tenants can receive desired communication environment without
deep knowledge about network and servers. An NS tenant decides
attribute values of its NS, such as bandwidth or latency, based on
their requirements, and NS providers design and create NSIs which
fulfill the values.
NS tenants need not to grasp detailed configurations in underlay
networks in this model. However, it may not be possible to provide
strictly desired NS to tenants because of abstruction of configurable
parameters. Moreover, it may cause complexity on designing NS
catalog due to quantities of selected attributes.
6.1.1. Capability in SaaS-like Model
In SaaS-like Model, an NS is represented for a tenant with attributes
values listed in Section 3.1. In other words, an NS tenant never
know the concrete configurations of components in underlay
infrastructure.
An NS tenant chooses a value from the range presented by the NS
provider in each attribute. The NS provider creates or changes a NS
by configuraing components in underlay infrastructures based on the
decided attribute values.
In terms of telemetry for assurance of service qualities on a NS, a
tenant can obtain telemetry information with unit of NSI, and never
know ones of underlay components structuring the NS.
6.2. PaaS-like Model
In PaaS-like Model, an NS is represented with several components such
as nodes and connectivities among them. An NS tenant can design and
customize its desired NS with combining such components. NS
providers breakdown the NS designed by the NS tenant to concrete
configurations of their infrastructure, and create/change NSSIs by
inputting the configurations. An NS tenant is also able to
incorporate its own functions or applications into its NSI by using
computing resources provided from NS providers.
This model potentially has high customizability of NS rather than
SaaS-like model, but needs NS tenants to have some knowledge about
network management. In terms of designing NS, the tenants provide
outline of their NSs, and thus it would make creation of concrete
configurations be easier.
Homma, et al. Expires January 9, 2020 [Page 16]
Internet-Draft draft-homma-slice-provision-models July 2019
6.2.1. Capability in PaaS-like Model
In PaaS-like model, an NS is represented with NF nodes and their
connectivities. An NS tenant can indicate functionalities of NF
nodes and thier locations. Also, the tenant decides attribute values
of connectivities. An NS provider creates or changes an NSI by
configuring underlay nodes and links depending on the indication of
the tenant. An NS tenant is also able to deploy its own NF as
software with provided computing resources.
In terms of telemetry, an NS tenant can obtain telemetry information
of NF nodes and connectivities structuring an NS, in addition to
whole of NSI.
6.3. IaaS-like Model
In IaaS-like model, an NS is represented with concrete configurations
of underlay infrastructure. NS tenants are able to structure or
change their desired NS by controlling infrastructure resources
directly.
This model potentially has high customizability of NS rather than
other models, but needs NS tenants to have deep knowledge about
network and server operation. Also, NS providers need not to
recognize NSs on their infrastructure because NS tenants directly
manage their NS. Meanwhile, in terms of security and prevention of
disturbances among NSs, some limitations on expositions of resources
to tenants would be needed.
6.3.1. Capability in IaaS-like Model
In IaaS-like Model, an NS is represented with configurations of
(virtual) nodes and (virtual) links connecting the nodes. An NS
tenant is able to configure nodes and links in underlay
infrastructure. In short, an NS tenant directly design detailed NS
and manages it. In addition, an NS tenant inserts its own functions
or applications in the NS with using computing resources.
In terms of telemetry, an NS tenant can obtain telemetry information
of nodes and links in addition of whole of NSI.
6.4. Mapping of NS Provision Models and Infrastructure Layering
An example of mapping of each NS provision model is shown in
Figure 3.
Homma, et al. Expires January 9, 2020 [Page 17]
Internet-Draft draft-homma-slice-provision-models July 2019
manage
[NS Tenant] ---------------------------+
|
|
. . . . . . . . . . . . . . . . . . . . . . . . |
*Service Layer |
.--. |
.------. ( )-. |
| Area#A |<==== Connectivity ====> .' [APL] ' SaaS-like|
`------' [BW:100Mbps, Delay<10ms]( ) <---------+
.------. ( [APL] -' |
| Area#B |<==== Connectivity ====> '-( ) |
`------' [BW:100Mbps, Delay<10ms] '---' |
Virtual Private |
Cloud |
. . . . . . . . . . . . . . . . . . . . . . . . |
*NS Layer |
__________________________________ |
/ / |
/ [AP]----o / PaaS-like|
/ / `--. /---[VM] / <---------+
/ [AP]---o-----o--[FW]--[VM] / |
/_________________________________/ |
NSI |
. . . . . . . . . . . . . . . . . . . . . . . . |
*Infra Layer |
|
|
[AP]----o /---[SV] |
/ `--. /---[SV] IaaS-like|
[AP]---o-----o--[FW]--[SV] <---------+
.---' /---[SV]
[AP]----'
*Legends
o : virtual/physical node
-- : virtual/physical link
[AP] : Access point
[APL]: Application owned by NS Tenant
[FW] : Firewall Function
[VM] : Virtual Machine Instance on Sever
[SV] : Server as Infrastructure
Figure 3: Mapping of NS provision models
Homma, et al. Expires January 9, 2020 [Page 18]
Internet-Draft draft-homma-slice-provision-models July 2019
In some cases, NSIs provided based on IaaS- or PaaS-like models are
coordinated to a form of SaaS-like model by an NS broker , and the NS
broker or by the tenant, becoming a NS provider in a recursive
manner. For example, a vertical customer sends its high-level
requirements to an NS broker create an appropriate NSI with resources
provided by infrastructure operators.
7. Security Considerations
In NSaaS, parts of controls of infrastructures are opened to
externals, and thus some mechanisms, such as authentication for APIs,
to prevent illegal access would be required.
Other considerations are TBD
8. IANA Considerations
This memo includes no request to IANA.
9. Acknowledgement
The author would like to thank Toru Okugawa for his kind review and
valuable feedback.
10. Informative References
[I-D.defoy-coms-subnet-interconnection]
Foy, X., Rahman, A., Galis, A.,
kiran.makhijani@huawei.com, k., and L. Qiang,
"Interconnecting (or Stitching) Network Slice Subnets",
draft-defoy-coms-subnet-interconnection-01 (work in
progress), October 2017.
[I-D.dong-teas-enhanced-vpn]
Dong, J., Bryant, S., Li, Z., Miyasaka, T., and Y. Lee, "A
Framework for Enhanced Virtual Private Networks (VPN+)
Service", draft-dong-teas-enhanced-vpn-03 (work in
progress), November 2018.
[I-D.homma-nfvrg-slice-gateway]
Homma, S., Foy, X., and A. Galis, "Gateway Function for
Network Slicing", draft-homma-nfvrg-slice-gateway-00 (work
in progress), July 2018.
[NGMN-5G-White-Paper]
NGMN, "NGMN 5G White Paper", February 2015,
<https://www.ngmn.org/5g-white-paper/5g-white-paper.html>.
Homma, et al. Expires January 9, 2020 [Page 19]
Internet-Draft draft-homma-slice-provision-models July 2019
[RFC8453] Ceccarelli, D., Ed. and Y. Lee, Ed., "Framework for
Abstraction and Control of TE Networks (ACTN)", RFC 8453,
DOI 10.17487/RFC8453, August 2018,
<https://www.rfc-editor.org/info/rfc8453>.
[TS.23.501-3GPP]
3rd Generation Partnership Project (3GPP), "3GPP TS 23.501
(V16.0.0): System Architecture for 5G System; Stage 2",
September 2018, <http://www.3gpp.org/ftp//Specs/
archive/23_series/23.501/23501-g00.zip>.
[TS.28.530-3GPP]
3rd Generation Partnership Project (3GPP), "3GPP TS 28.530
(V1.0.0): Management and orchestration of networks and
network slicing; Concepts, use cases and requirements
(work in progress)", June 2018,
<http://ftp.3gpp.org//Specs/
archive/28_series/28.530/28530-100.zip>.
[TS.28.541-3GPP]
3rd Generation Partnership Project (3GPP), "3GPP TS 28.541
(V15.1.0): 5G Netowkr Resource Model (NRM); Stage 2 and
stage 3 (Release 15)", June 2018,
<http://www.3gpp.org/ftp//Specs/
archive/28_series/28.541/28541-f01.zip>.
[TS.28.801-3GPP]
3rd Generation Partnership Project (3GPP), "3GPP TS 28.801
(V15.1.0): Study on Management and Orchestration of
Network Slicing for next generation network (Release 15)",
June 2018, <http://www.3gpp.org/ftp//Specs/
archive/28_series/28.801/28801-f10.zip>.
[WEBPUSH-WG]
IETF, "Web-Based Push Notifications(webpush)",
<https://datatracker.ietf.org/wg/webpush/about/>.
Appendix A. NS Structure in the 3GPP 5GS
The overview of structure of NS in the 3GPP 5GS is shown in Figure 4.
The terms are described in the 3GPP documents (e.g., [TS.23.501-3GPP]
and [TS.28.530-3GPP]).
Homma, et al. Expires January 9, 2020 [Page 20]
Internet-Draft draft-homma-slice-provision-models July 2019
<================== E2E-NSI =======================>
: : : : :
: : : : :
<====== RAN-NSSI ======><=TRN-NSSI=><====== CN-NSSI ======>VL[APL]
: : : : : : : : :
: : : : : : : : :
RW[NFs ]<=TRN-NSSI=>[NFs ]<=TRN-NSSI=>[NFs ]<=TRN-NSSI=>[NFs ]VL[APL]
. . . . . . . . . . . . .. . . . . . . . . . . . . ..
.,----. ,----. ,----.. ,----. .,----. ,----. ,----..
UE--|RAN |---| TN |---|RAN |---| TN |---|CN |---| TN |---|CN |--[APL]
.|NFs | `----' |NFs |. `----' .|NFs | `----' |NFs |.
.`----' `----'. .`----' `----'.
. . . . . . . . . . . . .. . . . . . . . . . . . . ..
RW RAN MBH CN DN
*Legends
UE: User Equipment
RAN: Radio Access Network
CN: Core Network
DN: Data Network
TN: Transport Network
MBH: Mobile Backhaul
RW: Radio Wave
NF: Network Function
APL: Application Server
Figure 4: Overview of Structure of NS in 3GPP 5GS
Authors' Addresses
Shunsuke Homma
NTT
Japan
Email: shunsuke.homma.fp@hco.ntt.co.jp
Hidetaka Nishihara
NTT
Japan
Email: nishihara.hidetaka@lab.ntt.co.jp
Homma, et al. Expires January 9, 2020 [Page 21]
Internet-Draft draft-homma-slice-provision-models July 2019
Takuya Miyasaka
KDDI Research
Japan
Email: ta-miyasaka@kddi-research.jp
Alex Galis
University College London
UK
Email: a.galis@ucl.ac.uk
Vishnu Ram OV
Independent Research Consultant India
India
Email: vishnu.u@ieee.org
Diego R. Lopez
Telefonica I+D
Spain
Email: diego.r.lopez@telefonica.com
Luis M. Contreras-Murillo
Telefonica I+D
Spain
Email: luismiguel.contrerasmurillo@telefonica.com
Jose A. Ordonez-Lucena
Telefonica I+D
Spain
Email: joseantonio.ordonezlucena@telefonica.com
Pedro Martinez-Julia
NICT
Japan
Email: pedro@nict.go.jp
Homma, et al. Expires January 9, 2020 [Page 22]
Internet-Draft draft-homma-slice-provision-models July 2019
Li Qiang
Huawei Technologies
China
Email: qiangli3@huawei.com
Reza Rokui
Nokia
Canada
Email: reza.rokui@nokia.com
Laurent Ciavaglia
Nokia
France
Email: Laurent.ciavaglia@nokia.com
Xavier de Foy
InterDigital Inc.
Canada
Email: Xavier.Defoy@InterDigital.com
Homma, et al. Expires January 9, 2020 [Page 23]