(19)

*EP004178169A1*
(11) EP 4 178 169 A1
(12) EUROPEAN PATENT APPLICATION

(43) Date of publication: (51) International Patent Classification (IPC):

10.05.2023 Bulletin 2023/19 H04L 41/0806 (2022.01) H04L 41/5054 (2022.01)
H04L 41/084 (2022.01) H04L 41/5041 (2022.01)
(21) Application number: 22204632.8 H04L 41/50 (2022.01) H04L 41/0273 (2022.01)
H04L 41/08 (2022.01) H04L 41/0895 (2022.01)
(22) Date of filing: 31.10.2022 H04L 41/0896 (2022.01) H04L 41/0897 (2022.01)
H04L 41/122 (2022.01) H04L 41/14 (2022.01)
H04L 41/40 (2022.01) H04L 43/0852 (2022.01)
H04L 43/20 (2022.01)

(52) Cooperative Patent Classification (CPC):

H04L 41/0843; H04L 41/0273; H04L 41/0806;
H04L 41/5048; H04L 41/5054; H04L 41/5083;
H04L 41/0886; H04L 41/0895; H04L 41/0896;
H04L 41/0897; H04L 41/122; H04L 41/145;
H04L 41/40; H04L 43/0858; H04L 43/20

(84) Designated Contracting States: (72) Inventors:

AL AT BE BG CH CY CZ DE DK EE ES FI FR GB • BUI, Dinh Thai
GR HR HU IE IS IT LI LT LU LV MC ME MK MT NL 92290 Chatenay Malabry (FR)
NO PL PT RO RS SE SI SK SM TR • BOLETTIERI, Simone
Designated Extension States: 56124 PISA (IT)
BA • BRUNO, RAFFAELE
Designated Validation States: 57128 LIVORNO (IT)
KH MA MD TN
(74) Representative: Novagraaf Technologies
(30) Priority: 04.11.2021 US 202163275913 P Bâtiment O2
2, rue Sarah Bernhardt
(71) Applicant: Nokia Solutions and Networks Oy CS90017
02610 Espoo (FI) 92665 Asnières-sur-Seine Cedex (FR)

(54) MULTI-ACCESS EDGE COMPUTING SLICING

(57) An apparatus comprises means for managing

an application service using a management web portal;
means for requesting, using the management web portal,
the application service with a given service level agree-
ment from a catalogue of offered application services;
and means for communicating, using the management
web portal, with an application service management
function, wherein the application service management
function is configured to translate the service level agree-
ment of the requested application service to a specifica-
tion of an application slice, and to trigger a creation of an
application slice instance with contacting an application
slice management function.
EP 4 178 169 A1

Processed by Luminess, 75001 PARIS (FR)

 1 EP 4 178 169 A1 2

Description network slice.

TECHNICAL FIELD FIG. 13 illustrates an alternative overlay deployment

model.
[0001] The examples and non-limiting embodiments 5
relate generally to communications and, more particular- FIG. 14 illustrates a stitching deployment mode,
ly, to multi-access edge computing slicing (MEC-SLICE). showing interconnection of the different slice sub-
nets to form the E2E application slice.
BACKGROUND
10 FIG. 15 is a block diagram of a multi-tenant MEC
[0002] It is known to implement partition functions of architecture supporting network and application slic-
logical entities in a communication network. ing.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 16 is an illustration of the orchestration entities

15 that are involved in the MEC domain.
[0003] The foregoing aspects and other features are
explained in the following description, taken in connec- FIG. 17 is an example apparatus configured to im-
tion with the accompanying drawings. plement the examples described herein.

FIG. 1 is a block diagram of one possible and non- 20 FIG. 18 is an example method performed with an
limiting system in which the example embodiments MEC-OO to implement the examples described
may be practiced. herein.

FIG. 2 is a block diagram of an ETSI reference ar- FIG. 19 is an example method performed with an
chitecture aligned with ESTI NFV. 25 MEC-CO to implement the examples described
herein.
FIG. 3 is a block diagram of an example MEC slicing
architecture. FIG. 20 is an example method performed with an
ASMF to implement the examples described herein.
FIG. 4 is a block diagram of an example MEC nested 30
slicing architecture. FIG. 21 is an example method performed with a UE
/ web portal to implement the examples described
FIG. 5 is a block diagram of an overall management herein.
architecture.
35 FIG. 22 is an example method performed with an
FIG. 6 is an exemplary illustration of the relationships APSMF to implement the examples described here-
between communication services, network slices, in.
network slice subnets and resources/network func-
tions. FIG. 23 is an example method performed with an
40 APSSMF to implement the examples described
FIG. 7 is a diagram depicting high level roles in a herein.
network slice framework.
DETAILED DESCRIPTION OF EXAMPLE EMBODI-
FIG. 8 depicts a 3GPP network slice management MENTS
architecture. 45
[0004] Turning to FIG. 1, this figure shows a block di-
FIG. 9 is a block diagram of an example MEC refer- agram of one possible and non-limiting example in which
ence architecture. the examples may be practiced. A user equipment (UE)
110, radio access network (RAN) node 170, and network
FIG. 10 is a diagram showing end-to-end network 50 element(s) 190 are illustrated. In the example of FIG. 1,
latency of a network slice instance (NSI) that in- the user equipment (UE) 110 is in wireless communica-
cludes MEC. tion with a wireless network 100. A UE is a wireless device
that can access the wireless network 100. The UE 110
FIG. 11 is a diagram showing high-level functional includes one or more processors 120, one or more mem-
roles in an MEC application slice framework. 55 ories 125, and one or more transceivers 130 intercon-
nected through one or more buses 127. Each of the one
FIG. 12 illustrates an overlay deployment model, or more transceivers 130 includes a receiver, Rx, 132
where the application slice is a consumer of an E2E and a transmitter, Tx, 133. The one or more buses 127

2
 3 EP 4 178 169 A1 4

may be address, data, or control buses, and may include Note that the DU 195 is considered to include the trans-
any interconnection mechanism, such as a series of lines ceiver 160, e.g., as part of a RU, but some examples of
on a motherboard or integrated circuit, fiber optics or oth- this may have the transceiver 160 as part of a separate
er optical communication equipment, and the like. The RU, e.g., under control of and connected to the DU 195.
one or more transceivers 130 are connected to one or 5 The RAN node 170 may also be an eNB (evolved NodeB)
more antennas 128. The one or more memories 125 in- base station, for LTE (long term evolution), or any other
clude computer program code 123. The UE 110 includes suitable base station or node.
a module 140, comprising one of or both parts 140-1 [0006] The RAN node 170 includes one or more proc-
and/or 140-2, which may be implemented in a number essors 152, one or more memories 155, one or more
of ways. The module 140 may be implemented in hard- 10 network interfaces (N/W I/F(s)) 161, and one or more
ware as module 140-1, such as being implemented as transceivers 160 interconnected through one or more
part of the one or more processors 120. The module buses 157. Each of the one or more transceivers 160
140-1 may be implemented also as an integrated circuit includes a receiver, Rx, 162 and a transmitter, Tx, 163.
or through other hardware such as a programmable gate The one or more transceivers 160 are connected to one
array. In another example, the module 140 may be im- 15 or more antennas 158. The one or more memories 155
plemented as module 140-2, which is implemented as include computer program code 153. The CU 196 may
computer program code 123 and is executed by the one include the processor(s) 152, memory(ies) 155, and net-
or more processors 120. For instance, the one or more work interfaces 161. Note that the DU 195 may also con-
memories 125 and the computer program code 123 may tain its own memory/memories and processor(s), and/or
be configured to, with the one or more processors 120, 20 other hardware, but these are not shown.
cause the user equipment 110 to perform one or more [0007] The RAN node 170 includes a module 150,
of the operations as described herein. The UE 110 com- comprising one of or both parts 150-1 and/or 150-2, which
municates with RAN node 170 via a wireless link 111. may be implemented in a number of ways. The module
[0005] The RAN node 170 in this example is a base 150 may be implemented in hardware as module 150-1,
station that provides access by wireless devices such as 25 such as being implemented as part of the one or more
the UE 110 to the wireless network 100. The RAN node processors 152. The module 150-1 may be implemented
170 may be, for example, a base station for 5G, also also as an integrated circuit or through other hardware
called New Radio (NR) . In 5G, the RAN node 170 may such as a programmable gate array. In another example,
be a NG-RAN node, which is defined as either a gNB or the module 150 may be implemented as module 150-2,
an ng-eNB. A gNB is a node providing NR user plane 30 which is implemented as computer program code 153
and control plane protocol terminations towards the UE, and is executed by the one or more processors 152. For
and connected via the NG interface (such as connection instance, the one or more memories 155 and the com-
131) to a 5GC (such as, for example, the network element puter program code 153 are configured to, with the one
(s) 190). The ng-eNB is a node providing E-UTRA user or more processors 152, cause the RAN node 170 to
plane and control plane protocol terminations towards 35 perform one or more of the operations as described here-
the UE, and connected via the NG interface (such as in. Note that the functionality of the module 150 may be
connection 131) to the 5GC. The NG-RAN node may distributed, such as being distributed between the DU
include multiple gNBs, which may also include a central 195 and the CU 196, or be implemented solely in the DU
unit (CU) (gNB-CU) 196 and distributed unit(s) (DUs) 195.
(gNB-DUs), of which DU 195 is shown. Note that the DU 40 [0008] The one or more network interfaces 161 com-
195 may include or be coupled to and control a radio unit municate over a network such as via the links 176 and
(RU). The gNB-CU 196 is a logical node hosting radio 131. Two or more gNBs 170 may communicate using,
resource control (RRC), SDAP and PDCP protocols of e.g., link 176. The link 176 may be wired or wireless or
the gNB or RRC and PDCP protocols of the en-gNB that both and may implement, for example, an Xn interface
control the operation of one or more gNB-DUs. The gNB- 45 for 5G, an X2 interface for LTE, or other suitable interface
CU 196 terminates the F1 interface connected with the for other standards.
gNB-DU 195. The F1 interface is illustrated as reference [0009] The one or more buses 157 may be address,
198, although reference 198 also illustrates a link be- data, or control buses, and may include any interconnec-
tween remote elements of the RAN node 170 and cen- tion mechanism, such as a series of lines on a mother-
tralized elements of the RAN node 170, such as between 50 board or integrated circuit, fiber optics or other optical
the gNB-CU 196 and the gNB-DU 195. The gNB-DU 195 communication equipment, wireless channels, and the
is a logical node hosting RLC, MAC and PHY layers of like. For example, the one or more transceivers 160 may
the gNB or en-gNB, and its operation is partly controlled be implemented as a remote radio head (RRH) 195 for
by gNB-CU 196. One gNB-CU 196 supports one or mul- LTE or a distributed unit (DU) 195 for gNB implementation
tiple cells. One cell may be supported with one gNB-DU 55 for 5G, with the other elements of the RAN node 170
195, or one cell may be supported/shared with multiple possibly being physically in a different location from the
DUs under RAN sharing. The gNB-DU 195 terminates RRH/DU 195, and the one or more buses 157 could be
the F1 interface 198 connected with the gNB-CU 196. implemented in part as, for example, fiber optic cable or

3
 5 EP 4 178 169 A1 6

other suitable network connection to connect the other [0013] The computer readable memories 125, 155,
elements (e.g., a central unit (CU), gNB-CU 196) of the and 171 may be of any type suitable to the local technical
RAN node 170 to the RRH/DU 195. Reference 198 also environment and may be implemented using any suitable
indicates those suitable network link(s). data storage technology, such as semiconductor based
[0010] It is noted that the description herein indicates 5 memory devices, flash memory, magnetic memory de-
that "cells" perform functions, but it should be clear that vices and systems, optical memory devices and systems,
equipment which forms the cell may perform the func- non-transitory memory, transitory memory, fixed memory
tions. The cell makes up part of a base station. That is, and removable memory. The computer readable mem-
there can be multiple cells per base station. For example, ories 125, 155, and 171 may be means for performing
there could be three cells for a single carrier frequency 10 storage functions. The processors 120, 152, and 175
and associated bandwidth, each cell covering one-third may be of any type suitable to the local technical envi-
of a 360 degree area so that the single base station’s ronment, and may include one or more of general pur-
coverage area covers an approximate oval or circle. Fur- pose computers, special purpose computers, microproc-
thermore, each cell can correspond to a single carrier essors, digital signal processors (DSPs) and processors
and a base station may use multiple carriers. So if there 15 based on a multi-core processor architecture, as non-
are three 120 degree cells per carrier and two carriers, limiting examples. The processors 120, 152, and 175
then the base station has a total of 6 cells. may be means for performing functions, such as control-
[0011] The wireless network 100 may include a net- ling the UE 110, RAN node 170, network element(s) 190,
work element or elements 190 that may include core net- and other functions as described herein.
work functionality, and which provides connectivity via a 20 [0014] In general, the various embodiments of the user
link or links 181 with a further network, such as a tele- equipment 110 can include, but are not limited to, cellular
phone network and/or a data communications network telephones such as smart phones, tablets, personal dig-
(e.g., the Internet). Such core network functionality for ital assistants (PDAs) having wireless communication
5G may include location management functions (LMF(s)) capabilities, portable computers having wireless commu-
and/or access and mobility management function(s) 25 nication capabilities, image capture devices such as dig-
(AMF(S)) and/or user plane functions (UPF(s)) and/or ital cameras having wireless communication capabilities,
session management function(s) (SMF(s)). Such core gaming devices having wireless communication capabil-
network functionality for LTE may include MME (Mobility ities, music storage and playback appliances having
Management Entity)/SGW (Serving Gateway) function- wireless communication capabilities, Internet appliances
ality. Such core network functionality may include SON 30 permitting wireless Internet access and browsing, tablets
(self-organizing/optimizing network) functionality. These with wireless communication capabilities, head mounted
are merely example functions that may be supported by displays such as those that implement virtual/augment-
the network element(s) 190, and note that both 5G and ed/mixed reality, as well as portable units or terminals
LTE functions might be supported. The RAN node 170 that incorporate combinations of such functions.
is coupled via a link 131 to the network element 190. The 35 [0015] UE 110, RAN node 170, and/or network ele-
link 131 may be implemented as, e.g., an NG interface ment (s) 190, (and associated memories, computer pro-
for 5G, or an S1 interface for LTE, or other suitable in- gram code and modules) may be configured to imple-
terface for other standards. The network element 190 ment (e.g. in part) the methods described herein, includ-
includes one or more processors 175, one or more mem- ing multi-access edge computing slicing. Thus, computer
ories 171, and one or more network interfaces (N/W 40 program code 123, module 140-1, module 140-2, and
I/F(s)) 180, interconnected through one or more buses other elements/features shown in FIG. 1 of UE 110 may
185. The one or more memories 171 include computer implement user equipment related aspects of the meth-
program code 173. ods described herein. Similarly, computer program code
[0012] The wireless network 100 may implement net- 153, module 150-1, module 150-2, and other ele-
work virtualization, which is the process of combining 45 ments/features shown in FIG. 1 of RAN node 170 may
hardware and software network resources and network implement gNB/TRP related aspects of the methods de-
functionality into a single, software-based administrative scribed herein. Computer program code 173 and other
entity, a virtual network. Network virtualization involves elements/features shown in FIG. 1 of network element(s)
platform virtualization, often combined with resource vir- 190 may be configured to implement network element
tualization. Network virtualization is categorized as either 50 related aspects of the methods described herein.
external, combining many networks, or parts of networks, [0016] Having thus introduced a suitable but non-lim-
into a virtual unit, or internal, providing network-like func- iting technical context for the practice of the example em-
tionality to software containers on a single system. Note bodiments, the example embodiments are now de-
that the virtualized entities that result from the network scribed with greater specificity.
virtualization are still implemented, at some level, using 55 [0017] Multi-access Edge Computing (MEC) allows for
hardware such as processors 152 or 175 and memories bringing cloud resources next to the end-user device (i.e.
155 and 171, and also such virtualized entities create user equipment) in order to meet low-latency require-
technical effects. ments of time-critical applications.

4
 7 EP 4 178 169 A1 8

[0018] ETSI MEC standard has specified a reference Owner entity setting up and deploying applications for all
architecture 200 aligned with ETSI NFV as illustrated in the different categories of applications. Without a direct
FIG. 2 extracted from ETSI MEC003 V2.2.1 (2020-12). access to the MEAO, this would mean the different MEC
The architecture includes MEAO 202 and OSS 204. Customers should provide these fine-grain settings to
[0019] The previous architecture 200 is not good 5 the MEC Owner via the aforementioned business proc-
enough as it does not allow for a multi-tenant MEC en- ess/channel which appears to hinder the deployment
vironment. More precisely, this architecture does not sup- speed.
port network slicing as per 3GPP definition allowing for [0022] Accordingly, the idea as described herein con-
an E2E (from the UE until the application is instantiated sists of splitting the MEAO 302 (referring also to MEAO
into an edge computing platform) network slice. Such a 10 202) into two with separate responsibilities. The new gen-
feature allows multiple actors for sharing the same MEC eral architecture is illustrated in FIG. 4.
infrastructure leading to cost optimization. [0023] The MEC (Infrastructure) Owner Orchestrator
[0020] A recent study (MEC Support for Network Slic- (MEC-OO) 402 is responsible for orchestrating and man-
ing: Status and Limitations from a Standardization View- aging the MEC Owner’s Network Slice Subnet dedicated
point - IEEE Communications Standards Magazine, June 15 to each tenant. The MEC Tenant/Customer Orchestrator
2020 - L. Comardi et al.), proposes to slice the MEC ar- (MEC-CO) 404 is responsible for orchestrating and man-
chitecture. The proposed architecture 300 is illustrated aging resources within the MEC Tenant’s Network Slice
in FIG. 3. Subnet. The MEC Tenant/Customer Orchestrator (MEC-
[0021] This architecture is still not good enough since CO) 404 could further partition those resources into
there is one single Operation Support system (OSS) 304 20 smaller slices called in this description as Application
and one single orchestrator (i.e. single MEAO 302) for Slices (APS - these could be also called as Network Slic-
multiple tenants. This architecture is not to be scalable es but Application Slices is used for clarity). A commu-
as both the single OSS 304 as well as the single MEAO nication interface 406 is established between the two Or-
302 have to manage connections to numerous devices chestrators (402, 404) in order to synchronize in terms
(e.g. UEs) pertaining to different tenants. For instance, 25 of resources regarding a Tenant’s Network Slice Subnet,
the single MEAO 302 should be aware of the tenant each but also in terms of forwarding rules (those managed by
UE 110 pertains to. This would create a bottleneck in the MEC-CO 404 and those managed by the MEC-OO
terms of customer facing when dealing with UE subscrip- 402).
tion addition/modification/deletion. The scalability issue [0024] In terms of management, each Tenant’s Net-
raised is not from a technical perspective (i.e. not related 30 work Slice Subnet can be managed by the Tenant’s NS-
to monolithic versus distributed implementations) but SMF 408 (Network Slice Subnet Management Function
rather from an operational, business and expertise per- - as per 3GPP definition). Similarly, each Tenant’s Ap-
spective. From an operational/business perspective, the plication Slice can be managed by the Tenant’s new
MEAO is not sliced or partitioned within the state-of-the- building block called Application Slice Management
art. It appears to be mono-tenant. This tenant may be 35 Function (APSMF) which delegates the life-cycle man-
called the MEC Owner. In such a situation, other tenants agement of constituent Application Slice Subnets
(referred to as MEC Customers) cannot directly provision (APSS) to a different domain Application Slice Subnet
the (MEC Owner’s) MEAO with their respective UE in- management Function (APSSMF). The latter is also a
formation and related policies. The MEC customers must new management building block to manage an Applica-
communicate the UE information and related policies to 40 tion Component Function (ACF).
the MEC Owner via a business channel/process which [0025] The MEC reference architecture is split into two
is ruled by agreed contracts. The MEC Owner needs to responsibility domains with one of them managed and
validate those items of information (with regards to con- orchestrated by the MEC Owner 414 and the other by
tracts) before enforcing them onto the MEAO. It appears the MEC Customer 416. The MEC Owner 414 plays both
therefore as if the ’business interface’ of the MEAO was 45 Hardware supplier and NFVI supplier roles (with refer-
a bottleneck. From an expertise perspective, a tenant or ence to 412 and possibly 410). The MEC Owner 414
a MEC Customer whose role is to provide Application provides MEC network slices to the MEC Tenant also
Services should have the expertise necessary to read referred to as the MEC Customer 416. The MEC Cus-
the user manual of the different ACFs to deploy. This is tomer 416 is a Tenant with at least a dedicated MEC
similar to the role of a network operator who should have 50 network slice. The MEC Customer 416 plays the role of
the required expertise to read the user manuals of net- a MEC operator with all standard MEC components and
work functions being deployed in his/her networks. How- interfaces as per ETSI MEC003.
ever, application categories are tremendously larger than [0026] This dichotomy allows a single MEC Owner’s
network function types. Some application categories may infrastructure to host multiple MEC Customers (i.e. mul-
require specific expertise to integrate and to deploy. For 55 tiple tenants) with each of them having its own MEC net-
instance, some applications for health professionals re- work slice (i.e. its dedicated data plane (424, 425) pro-
quire some minimum health education to set them up vided by the MEC Owner 414) with related management
correctly. It is therefore difficult to imagine a single MEC capability. In turn, each MEC Customer 416 manages

5
 9 EP 4 178 169 A1 10

and orchestrates his/her own MEC application slices. delegate the 5GC traffic influence (i.e. enforce new 5GC
[0027] Implementation-wise, this is made possible us- forwarding rules or tear down old ones) to the MENSM
ing the nested virtualization capability of the infrastruc- 428 either via a direct link or via the MEC-CO-to-MEC-
ture. For instance, the MEC Customer 416 can deploy OO link 406.
Virtual Machines (VMs) within VMs provided by the MEC 5 [0032] Among the new building blocks identified in-
Owner 414. Alternatively, the MEC Customer 416 could clude the MEC Customer Orchestrator 404, the MEC
deploy containers (e.g. Docker containers) within the lat- Owner Orchestrator 402, the MEC Network Slice man-
ter VMs. ager 428, an Interface 406 between the MEC Customer
[0028] The MEC Owner Orchestrator (MEC-OO) 402 Orchestrator 404 and the MEC Owner Orchestrator 402,
receives from the MEC NSSMF 408 (see "Management 10 and an Interface between the (MEC Customer) MEC
Hierarchy" sub-section below for more detail) orders to Platform Manager 434 and the (MEC Owner) MEC Net-
create, modify or delete MEC network slices. The MEC work Slice Manager 428, possibly via the previously listed
Owner Orchestrator (MEC-OO) 402 collaborates with the interface 406, namely the Interface 406 between the
MEC Owner VEO 426 (MEC-OVEO - e.g. OpenStack MEC Customer Orchestrator 404 and the MEC Owner
Heat, NFVO) to manage such network slice life-cycle, 15 Orchestrator 402.
which may be performed together with MEC network slice [0033] In FIG. 4, items within the MEC Owner domain
LCM 442. For simplicity, this description considers that 414 include MEC Owner Orchestrator 402, VEO 426,
the MEC Customer 416 is provided a dedicated network MEC network slice manager 418, MEC network slice
slice as per {Cominardi2020}. For instance, the MEC LCM 442, virtual infrastructure manager 440, and NFVI
Owner 414 provides to the MEC Customer 416 with a 20 412.
dedicated Kubernetes cluster connected to the 5G Core [0034] In FIG. 4, items within the MEC customer do-
(5GC) via a dedicated data plane slice. The MEC-OO main 416 include User app LCM proxy 450, MEC cus-
402 also collaborates with the MEC Network Slice Man- tomer orchestrator 404, VEO 432, MEC application slice
ager 428 to manage the network slice data plane param- #1 418, MEC application slice #2 420, MEC application
eters. For instance, the MEC Network Slice Manager 428 25 slice #3 422, MEC platform 436, MEC platform manager
could behave like a 3GPP Application Function (AF) 434, MEC platform LCM 444, MEC application LCM 446,
which interacts with the 5GC to synchronize data plane MEC application slice LCM 448, virtual infrastructure
forwarding rules to realize local breakout regarding MEC manager 438, NFVI 410, data plane 424 within NFVI 410,
application traffics. and data plane 425 within NFVI 412.
[0029] The MEC Customer Orchestrator (MEC-CO 30 [0035] FIG. 4 also shows that data plane 424 within
404) receives from the MEC APSSMF 430 (see "Man- NFVI 410 comprises an instance of each of the MEC
agement Hierarchy" sub-section below for more detail) application slice #1 418, MEC application slice #2 420,
orders to create, modify or delete MEC application slices. and MEC application slice #3, 422. Also shown in CFS
The MEC Customer Orchestrator (MEC-CO 404) collab- portal 452 and device application 454. One or multiple
orates with the MEC Customer VEO 432 (MEC-CVEO - 35 application slices can be provided by the MEC Customer
e.g. K8S) to manage such MEC application slice life-cy- as an ASP (Application Service Provider) to an ASC (Ap-
cle, which may be performed together with MEC platform plication Service Consumer).
LCM 444, MEC app LCM 446, and/or MEC app slice LCM
448. The MEC-CO 404 via the MEC-CVEO 432 provides Management hierarchy:
compute, storage and network resources to each MEC 40
application slice. The MEC-CO 404 also collaborates [0036] Definitions include the following for NF, VNF,
with the MEC Customer Platform Manager (MECPM) 434 and the application component function (ACF).
to manage an ETSI MEC Platform 436 instance (e.g. [0037] NF (3GPP definition): "a functional block within
embodied as a Docker container). a network infrastructure that has well-defined external
[0030] In order to stitch the MEC application slice data 45 interfaces and well-defined functional behavior."
plane to the MEC network slice data plane, the MEC-OO [0038] VNF (3GPP definition): "an implementation of
402 creates gateways that it communicates to the MEC- a Network Function using virtualization techniques of-
CO 404. fered by the Network Function Virtualization Infrastruc-
[0031] Regarding interaction with the 5GC, each MEC ture (NFVI)."
Platform 436 (MEP - within the MEC Customer’s respon- 50 [0039] Application Component Function (ACF): is the
sibility domain 416) could play the role of a 3GPP Appli- atomic data-processing component of an (MEC) appli-
cation Function (AF) and interact with the 5GC to influ- cation which presents well-defined external interfaces
ence 5G traffic from/and towards MEC applications. Al- and Application Programmable Interfaces (APIs) and
ternatively, the MEC Owner’s MEC Network Slice Man- well-defined functional behavior (e.g. Docker container).
ager (MENSM) 428 can play the role of an AF. The MEC 55 [0040] Since VNF and ACF are designed with different
Network Slice Manager 428 could interact with the 5GC purposes, it is not sensible to use 3GPP-defined CSMF
to influence 5G traffic from/to MEC applications on behalf to manage Application Services (ASs) built from ACFs.
of the MEC Customer. In this case, the MEP 436 could Indeed, 3GPP-defined CSMF is essentially specified to

6
 11 EP 4 178 169 A1 12

manage Communication Services (CSs) which do not mentioned E2E AS SLA splitting. In this stitching model,
have the same SLA framework as ASs. For instance, the APSMF 506 is responsible to communicate directly
data backup is the main concern for SLA within AS con- with the NSMF 510 to manage the network slice subnets
text while it is not part of CS SLA. associated with the APS.
[0041] Similarly to CSs, ASs can be nested. For in- 5 [0048] Finally, the management of network slice via
stance, an Application Consumer (ASC) A uses face rec- the NSMF 510 and the one of per-domain network slice
ognition Application Service from Application Service subnets via NSSMFs are well-defined by 3GPP, and they
Provider (ASP) B. It adds ACFs which retrieve video do not need to be extended (Cf. 3GPP TS 28.531, 3GPP
stream from customers’ cameras and ACFs which take TS 28.541). Also shown in FIG. 5 is the RAN NSSMF
control of customers’ doors to build a new AS which pro- 10 514, the core NSSMF 516, the TN NSSMF 518, and the
vides automatic door opening based on face recognition. MEC NSSMF 520.
ASC A then becomes ASP A selling the previous new [0049] Among the new building blocks identified in-
AS to customers such as house/apartment rental plat- clude the ASMF 504, the APSMF 506, the APSSMFs
forms. 512 and 522, the Interface 528 between ASMF 504 and
[0042] Similarly to a Network Slice built upon a set of 15 CSMF 508, the Interface 526 between ASMF 504 and
VNFs, this document coins the concept of Application APSMF 506, and the Interface 530, 532 between APSMF
Slice (APS) which is built upon a non-empty set of ACFs 506 and APSSMF (512, 522).
and, possibly, on or more VNFs. [0050] An advantage and technical effect of the exam-
[0043] By extending the 3GPP overall management ar- ples described herein include that the general architec-
chitecture to an Application Slice, it is possible to have 20 ture allows for separating/isolating tenant-dedicated
an overall management architecture as shown in FIG. 5. MEC orchestration work. This leads to a more scalable
[0044] A user can manage its AS and possibly related orchestration solution. A further advantage and technical
NSs via a management Web Portal 524. The user can effect of the examples described herein include that the
request an application service (AS) with a given SLA from architecture supports hierarchical MEC slicing as an
a catalogue of offered ASs. The business interaction of 25 MEC Owner (provider of physical servers) that can slice
the web portal 524 can happen in two manners depend- its infrastructure to serve multiple independent MEC Cus-
ing on the AS/APS deployment models that is used in tomers, who in turn can independently orchestrate allo-
the system. In both cases, the web portal 524 communi- cated resources to their tenants. Nested virtualization
cates with a new management function, called Applica- technologies can enable such a nested-slicing architec-
tion Service Management Function (ASMF) 504, which 30 ture. A further advantage and technical effect of the ex-
is responsible for translating the SLA of the requested amples described herein is providing an associated over-
AS to the specification of an application slice (APS) and all management architecture.
to trigger the creation of the APS instance by contacting [0051] Within an MEC infrastructure, the examples de-
a new management function called Application Slice scribed herein may be implemented with two MEC Or-
Management Function (APSMF) 506. The APSMF 506 35 chestrators belonging to different administration domains
splits the APS into multiple subnets, one for each domain with one domain (MEC Owner) providing MEC hosts to
over which the requested APS spans, including possibly the other (MEC Customer).
the network and the AS’s endpoints, namely the UE re- [0052] Accordingly, the examples described herein re-
questing the AS, and the edge system instantiating the late to multi-access edge computing application slice ar-
AS. To this end, introduced herein is the new Application 40 chitecture and proof-of-concept.
Slice Subnet Management Function (APSSMF), which [0053] Amongst different design focuses, 5G stand-
applies the APSS life-cycle management commands ards have worked on low-latency requirements, which
within the two potential domains that are relevant for an work caters to safety-critical and mission-critical applica-
application slice subnets, namely the UE 512 and the tions. Industrial automation within Industry 4.0 is an ex-
MEC 522. 45 ample of new perspectives that are triggered by such a
[0045] In the overlay model (label 1 501), the E2E AS- new capability. To reach low-latency targets, the 5G New
MF 504 is also responsible for translating the E2E AS Radio (NR) standard has defined the Ultra-Reliable Low
SLA into E2E CS SLA and for requiring the adapted E2E Latency Communication (URLLC) which allows for
CS from the CSMF 508. reaching a few ms one-way latency with small radio
[0046] In one of the stitching models (label 2 502), the 50 frames. To further improve end-to-end latency, the 5G
web portal 524 (expert user) is also responsible for break- architecture has integrated distributed cloud infrastruc-
ing the E2E AS SLA into two separate and independent tures, namely the Multi-access Edge Computing (MEC),
SLAs. The first is related to the E2E CS that the E2E which allows for bringing cloud resources next to the end-
ASMF 504 requests from the CSMF 508. The second user device.
SLA is restricted to the AS endpoints, namely the UE and 55 [0054] Another important pillar of the 5G standards is
the MEC, and we can refer to it as EP AS SLA. the network slice concept which allows for sharing the
[0047] Alternatively (label 3 503), the web portal 524 same 5G infrastructures amongst different usages. This
(non-expert user) does not need to perform the afore- aims at maximizing 5G operator revenue. An end-to-end

7
 13 EP 4 178 169 A1 14

network slice framework taking into account the transport ture. Current trends to integrate support for network slic-
network has been proposed by [ETSI, Zero-touch net- ing into the MEC standard are further described herein,
work and service management (zsm); end-to-end man- as well as identification of gaps to be filled to extend slic-
agement and orchestration of network slicing, Group ing to the edge, which such gaps are filled by the exam-
Specification ETSI GS ZSM 003 V1.1.1, ETSI ISG ZSM 5 ples described herein.
(06 2021)] . There are several definitions of a network
slice, but it has been agreed that a network slice is a 3GPP Network Slice Concept and Building Blocks
logical network composed of dedicated and shared net-
work resources and that there is isolation between net- [0058] The 3GPP approach to network slicing is inher-
work slice instances at different levels namely at the data 10 ited from the NGMN network slicing concept, and it is
plane, the control plane and the management plane. based on the distinction between network slices and net-
[0055] As latency should be managed end-to-end, work slice instances (NSIs). Specifically, a network slice
there is a need to establish and manage an end-to-end is defined as "a logical network that provides specific net-
network slice including the MEC. There are also propos- work capabilities and network characteristics, supporting
als to start such an integration. However, still missing is 15 various service properties for network slice customers"
a complete architecture which allows for taking into ac- [refer to 3GPP, Management and orchestration; con-
count processing and storage latency within the MEC. cepts, use cases and requirements (release 17), Tech-
This architecture should also include the overall slice nical Specification 28.530 V17.1.0, (March 2021)], while
management building block. a NSI is an activated network slice, namely "...a set of
[0056] The examples described herein fill those gaps. 20 network functions and the resources for these network
The examples described herein provide for four main functions which are arranged and configured ...to meet
technical effects. First, introduced is a new concept of certain network characteristics..." that are required by a
slice which allows for better distinguishing between ap- communication service [refer to 3GPP, Telecommunica-
plication and network services. Second, described is an tion management; Study on management and orches-
end-to-end slice architecture which allows for taking into 25 tration of network slicing for next generation network (Re-
account processing and storage latencies, especially lease 15), Technical Report 28.801 V15.1.0, (Jan.
those within the MEC. Third, described is an overall man- 2018).]. Following the relationship between CS and NS,
agement architecture which allows for operating the life- the 3GPP network slicing architecture is organized into
cycle management of such end-to-end slices. Finally, de- three distinct logical layers: 1) a service instance layer,
scribed is an implementation of the described architec- 30 2) a network slice instance layer, and 3) a resource layer.
ture while focusing on the MEC domain. The first layer encompasses the service instances (a.k.a.
Communication services (CS) in 5G networks) that are
Network Slicing associated with service-level agreements (SLA), namely
business contracts between the service provider and the
[0057] Early network designers proposed to set up vir- 35 clients, which specify the service levels to be ensured.
tual LANs to provide network segmentation at the data The second layer comprises the NSIs that are deployed
link layer, addressing issues such as scalability, security, to serve the CS requirements. Each NSI is associated to
and network management of switched Ethernet net- a Network Slice Template (NST), which describes the
works. However, over the years, the concept of network general structure and configurations of the network slice,
slicing has significantly evolved. One of the main drivers 40 and the Service Level Specification (SLS), which lists the
of this technology has become the need for network serv- set of technical attributes that have to be satisfied by the
ice flexibility and programmability to efficiently accom- deployed network slice. In other words, the SLS trans-
modate diverse business use cases. More recently, a lates the business objectives of the SLA to network char-
network slice is commonly defined as a logical network acteristics. Finally, the third layer includes the necessary
that provides specific network capabilities and charac- 45 physical (hardware and software) and logical resources
teristics to support certain communication services serv- to support the NSIs.
ing a business purpose while ensuring functional and per- [0059] Although a network slice is designed from an
formance isolation. While there is a consolidated high- end-to-end perspective, the 3GPP management archi-
level view of network slicing, various network slicing mod- tecture recognizes that a network slice spans different
els are currently under definition depending on the virtu- 50 technical domains, namely device, access network, core
alization technology, the network architecture, and the network, transport network and network management
standard development organization (SDO). Refer to [ET- system, with separate scope and technologies. For this
SI MEC ISG, Multi-access Edge Computing (MEC); Sup- reason, the 3GPP network slicing architecture introduces
port for network slicing, Group Report MEC 024 V2.1.1, the concept of Network Slice Subnets (NSSs), defined
ETSI (Nov. 2019)] for a comprehensive comparison of 55 as "...a set of network functions and the associated re-
different slicing concepts. The 3GPP approach to net- sources (e.g. compute, storage and networking resourc-
work slicing provides the foundation for the herein de- es) supporting network slice." [refer to 3GPP, Manage-
scribed end-to-end (E2E) slicing management architec- ment and orchestration; concepts, use cases and re-

8
 15 EP 4 178 169 A1 16

quirements (release 17), Technical Specification 28.530 sumes communication services.

V17.1.0, (March 2021)]. If a 3GPP CS can be seen as a
business wrapper of one or more network slices, in turn, • Communication Service Provider (CSP) 704: pro-
a network slice is a wrapper of one or more network slice vides communication services that are designed,
subnets with some Service Level Specification (SLS). 5 built and operated with (or without) a network slice.
FIG. 6 illustrates the relationship between these different
concepts. In FIG. 6, there are several points worth noting. • Network Operator (NOP) 706: designs, builds and
Network Functions (NF) refer to processing functions of operates network slices.
the 5G network (both access 604 and core networks 602),
which expose APIs to provide one or more services to 10 • Network Equipment Provider (NEP) 708: supplies
other NFs, following the producer-consumer concept. network equipment including VNFs to a network op-
NFs include physical network nodes (namely Physical erator.
Network Functions or PNF) and Virtual Network Func-
tions (VNF). A VNF is a software implementation of a • Virtualization Infrastructure Service Provider (VISP)
network function within an NFV infrastructure (NFVI). In 15 710: provides virtualized infrastructure services.
FIG. 6, network slice subnet AN-1 606 and network slice
subnet AN-2 608 each contain distinct sets of NFs of the • Data Centre Service Provider (DCSP) 714: provides
access network (610, 612, 614, 616), while network slice data centre services.
subnet CN-2 620 and network slice subnet CN-3 622
partly share some of the NFs of the core network, includ- 20 • NFVI Supplier 712: supplies a network function vir-
ing shared NF 632. tualization infrastructure.
[0060] The network operator offers network slice sub-
net 1 638 as a network slice 1 644 to CS A 650. For this • Hardware Supplier 716: supplies hardware.
purpose, the network operator associates the SLS de-
rived from the SLA of CS A 650 to network slice subnet 25 [0063] An organization can play one or several roles
1 638. Note that a network slice can satisfy the service simultaneously (for example, a company can play CSP
requirements of different communication services 660 and NOP roles simultaneously). Further, as shown in
(e.g., network slice 1 644 in FIG. 6 is serving both CS A FIG. 7, the entities function as a client, provider, or both
650 and CS B 652). Finally, it is helpful to point out that a client and provider.
the deployment template of a network slice subnet should 30 [0064] The 3GPP has also standardized the orches-
contain the descriptors of its constituent NFs and infor- tration and management functions for the life-cycle man-
mation relevant to the links between these NFs, such as agement of network slices. Specifically, the 3GPP slicing
the topology of connections and QoS attributes for these management architecture is reduced to three essential
links (e.g. bandwidth). The latter information can be rep- network functions, called CSMF 804, NSMF 806, and
resented in the form of a graph, using the data information 35 NSSMF (made up of RAN NSSMF 808, Core NSSMF
model defined by the ETSI NFV standard for describing 810, and TN NSSMF 812), as also illustrated by FIG. 8.
the Virtualized Network Function Forwarding Graph Web portal 802 provides a management function to ac-
(VNFFG). cess the various slice related management functions.
[0061] FIG. 6 further shows a transportation network [0065] The Communication Service Management
670. FIG. 6 also shows communication service C 654, 40 Function (CSMF) 804 is the user interface for slice man-
network slice #2 646, network slice #3 648, network slice agement. It converts the SLAs of the CSs requested by
subnet #2 640, network slice subnet # 3 642, NSS CN- the CSC into SLS and delegates the management of
1 618 comprising NFs 624 and 626, NFs 628 and 630 of NSIs to NSMFs. The Network Slice Management Func-
network slice subnet #2 640 which are not shared with tion (NSMF) 806 manages NSIs and splits them into sub-
network slice subnet #3 642, and NFs 634 and 636 of 45 nets for the RAN, transport and core domains. Then, the
network slice subnet #3 642 which are not shared with NSMF 806 delegates the management of slice subnets
network slice subnet #2 640. to NSSMFs (808, 810, 812). The Network Slice Subnet
Management Function (NSSMF) (808, 810, 812) applies
3GPP Network Slice Management Architecture the NSMF’s life-cycle management commands (e.g., in-
50 stantiate, scale, terminate, remove) within a particular
[0062] In order to lay the groundwork for the design of subnet.
an overall network slice management architecture, the [0066] 3GPP also assumes that a CSC can rely on a
3GPP has defined high-level operational roles, which web portal to request the creation of a network slice, using
permit to draw clear boundaries in terms of operational the NST made available by the CSMF 804. It is also worth
responsibilities. FIG. 7 illustrates the different roles as 55 pointing out that the NSSMF (808, 810, 812) is where
identified by the 3GPP, which are defined as follows: most of the slice intelligence resides. It takes a command
from the NSMF, such as "build a slice," and activates it
• Communication Service Customer (CSC) 702: con- by doing all the behind-the-scenes work of function se-

9
 17 EP 4 178 169 A1 18

lection, storage, configuration, and communication. ations Support System (OSS) 924 of the network oper-
Once each sub-slice is created, the NSMF 806 is in ator, which receives the requests for instantiation or ter-
charge of stitching them together to build the end-to-end mination of applications from either applications 926 run-
network slice. ning in the devices (e.g. UEs (110, 928) or from third-
5 party customers 930 through the CFS portal 932. It is
ETSI Multi-access Edge Computing (MEC) helpful to point out that the MEC standard provides a
complete specification of only a limited set of necessary
[0067] The ETSI organization has introduced the Multi- APIs, and associated data models and formats, but most
access Edge Computing (MEC) since 2014 to provide a of the management reference points are voluntarily left
standard framework for the development of inter-opera- 10 open by the standard to foster market differentiation.
ble applications over multi-vendor edge computing plat-
forms. To this end, the MEC technology provides a new MEC and Network Slicing
distributed software development model containing func-
tional entities, services, and APIs, enabling applications [0069] As pointed out, the MEC is a distributed com-
to run on top of a generic virtualization infrastructure lo- 15 puting environment at the edge of the network, on which
cated in or close to the network edge. For the sake of multiple applications can be served simultaneously while
discussion, FIG. 9 shows the generic ETSI MEC refer- ensuring ultra-low latency and high bandwidth. To
ence architecture 900, which consists of three main achieve this goal, applications have real-time access to
blocks: (i) the MEC Host 902, (ii) the MEC Platform Man- network information through APIs exposed by MEC serv-
ager (MEPM) 904 and (iii) the MEC Orchestrator (MEO) 20 ices. According to the MEC standard [refer to ETSI, Multi-
906. access edge computing (mec); framework and reference
[0068] The MEC host 902 is at the core of the MEC architecture, Group Specification ETSI GS MEC 003
architecture 900 as it contains: (i) the generic virtualiza- V2.2.1, ETSI ISG MEC (12 2020)], each MEC system is
tion infrastructure (VI) 908, which provides compute, stor- mandated to offer services providing authorized applica-
age, and network resources for the MEC applications; 25 tions with (i) radio network information (such as network
(ii) the MEC applications 910 running on top of the VI 908 load and status), and (ii) location information about UEs
(existing ETSI MEC specifications assume that MEC ap- served by the radio node(s) associated with a MEC host.
plications 910 are deployed as VMs using a hypervisor- Furthermore, the Bandwidth Manager service, when
based virtualization platform, but alternative virtualization available, permits both the allocation of bandwidth to cer-
technologies and paradigms are under consideration); 30 tain traffic routed to and from MEC applications and the
and (iii) the MEC platform 912, an environment that hosts prioritization of that traffic, also based on traffic rules re-
MEC services 914 and offers to authorized MEC appli- quired by applications. Based on the above, it is straight-
cations 910 a reference point to discover and consume forward to realize that the fundamental design behind the
MEC services 914, as well as to announce and offer new MEC architecture is to enable a network-aware applica-
MEC services 914. As discussed in detail further herein, 35 tion design, namely, to allow MEC applications and MEC
MEC services 914 are an essential component of a MEC platforms to leverage network information to satisfy their
system, as they allow MEC applications 910 to be net- requirements.
work-aware. In addition, the MEC platform 912 is respon- [0070] On the contrary, the 3GPP-based network slice
sible for configuring 916 a local DNS server and instruct- concept envisions an architectural shift as it relies on a
ing 918 the data plane of the VI on how to route traffic 40 communication service-centric network provisioning. For
among applications, services, existing DNS serv- this reason, the ETSI MEC group has recently started
ers/proxies, and external networks. A management layer discussing which new MEC functionalities and interfac-
920 is associated with the MEC hosts 902 of a MEC sys- es, as well as extensions to existing MEC components,
tem, including a virtualization infrastructure manager are required to support network slicing, e.g. by including
(VIM) 922 and a MEC platform manager (MEPM) 904. 45 network slice ID into different MEC interfaces. The ETSI
The VIM 922 is responsible for allocating, maintaining, report has identified several use cases based on the dif-
and releasing the virtual resources of the VI 908. The ferent network slicing concepts that are advocated in dif-
MEPM 904 is responsible for managing the life-cycle of ferent SDOs.
MEC applications 910 and informing individual MEC plat- [0071] The most advanced proposal for supporting
forms 912 about application rules, traffic rules 918, and 50 network slicing in MEC systems is the so-called MEC-in-
DNS configuration 916. Finally, the MEO 906 is the core NFV architecture which permits to deploy MEC applica-
functionality of the MEC system-level management. Spe- tions and MEC platforms as VNFs. To integrate MEC
cifically, the MEO 906 is responsible (i) for selecting the hosts in a NFV environment, the ETSI standard proposes
MEC host(s) 902 for application instantiation based on to substitute (i) the MEPM with a MEPM-V entity that
application requirements and constraints (e.g. latency), 55 delegates the MEC-based VNF life-cycle management
available resources, and available services; and (ii) for to one or more VNF managers (VNFM), and (ii) the MEO
maintaining an updated view of the MEC system 900. with a MEC application orchestrator (MEAO) that dele-
Furthermore, the MEO 906 is interfaced with the Oper- gates to the NFV orchestrator (NFVO) the resource or-

10
 19 EP 4 178 169 A1 20

chestration for MEC VNFs. Different uses cases are also sources to the served customers.
possible in the context of the MEC-in-NFV architecture,
for instance, to enable the sharing of a MEC host with E2E Slice Architecture for MEC Applications
several NSIs, or to allow MEC applications belonging to
multiple tenants to be deployed in a single NSI. ETSI 5 [0075] Described herein are the concepts of applica-
MEC also recognizes the importance of the contribution tion services (AS) and application slices (APS), and how
of MEC applications to end-to-end latency. Thus, a use these concepts are instantiated in the context of MEC
case exists in which the MEC platform is included in a systems. An application slice is described herein from
3GPP NSI, and the end-to-end network delay budget both operation and business perspectives with introduc-
takes into account MEC components and in particular, 10 ing operation and business roles, respectively, and rela-
the delay from the data network to the MEC platform as tionships of the different entities involved in application
shown in FIG. 10. slice and service provisioning. Also described herein are
[0072] In particular, FIG. 10 shows access network la- the different architectural use cases that correspond to
tency 1001 between the UE 1020 and the network node provided model. Further described herein is an orches-
1030, network node processing latency 1002, transport 15 tration/management architecture for 3GPP networks that
network latency 1003 between the network node 1030 enable such new use cases. An extended MEC reference
and the UPF 1040, UPF processing latency 1004, and architecture required to support the envisioned manage-
internet network latency 1005 between the UPF 1040 ment architecture is described.
and the MEC host 1050. FIG. 10 also shows the control
plane 1010 coupled to the network node 1030 and the 20 Preliminary Concepts
UPF 1040.
[0073] In this case, the descriptor of the network serv- [0076] As may be inferred from the description herein,
ices to be created in the NFV environment shall include the 3GPP CSs cannot be straightforwardly used to model
the network latency requirement, which is distributed to the services offered by an edge computing platform. For
the access network, core network, transport network and 25 the sake of disambiguation, in this description, the serv-
data network. This would require extending the data mod- ices offered to the customers of an edge computing plat-
el of the NFV service descriptors to include the Applica- form are referred to as Application Services (ASs). The
tion Descriptor, which contains fields to indicate the type main reason is that CSs and ASs are designed to support
of traffic to offload and the MEC service to consume. different business purposes, i.e., regulated by different
However, such a proposal has the drawback of requiring 30 SLA frameworks. Specifically, CSs are dedicated to the
the 5G CSMF to translate application-service-related re- transmission of data and network traffic, while ASs are
quirements into network-slice-related requirements. In- designed to support services which are generally not
deed, SLA frameworks regarding application services communication oriented in essence, but focus on data
are typically different from communication services (as processing, such as Augmented Reality (AR), Virtual Re-
per the 3GPP definition). For instance, data backup is 35 ality (VR), UltraHigh Definition (UHD) video, real-time
often part of an application service SLA but is not part of control, etc. From the management perspective, it would
a communication service SLA. Thus, radical changes in not be sensible to manage ASs with the existing 3GPP
CSMF implementations would be needed to support such CSMF. Some extensions or new management functions
use cases. should be added to the overall management framework
[0074] A harmonized view of the possible use cases 40 to translate application-related SLAs into application-re-
may be defined in to enable MEC frameworks to provide lated SLSs and network-related SLSs.
transparent end-to-end multi-slice and multi-tenant sup- [0077] In analogy to the 3GPP network slice concept,
port by facilitating the sharing of MEC applications and an Application Slice (APS) is defined as a virtualized data
MEC platforms among multiple NSIs. To this end, an in- collection and a set of data processing resources that
ter-slice communication mechanism is described that au- 45 provides specific processing functions with adapted SLS
tomatically filters exchanged data between slices in- to support an Application Service and the associated
stalled on the same MEC facilities. Focusing on 5G net- SLA. The APS can possibly include Network Functions
works, extensions of the slice management architecture to ensure correct data communication among the previ-
can enable the integration of MEC as a sub-slice by in- ous processing functions. An aspect of the application
troducing the MEC NSSMF component. The examples 50 slice concept is the support of isolation not only of the
described herein implements a slice management archi- resources that are used by each application slice but also
tecture as one of the starting points of an envisioned busi- of the management data and functions. Similarly to a
ness model in which multiple tenants exist who are willing network slice, the application slice is an end-to-end con-
to pay to get isolated slices of network and edge com- cept and can span multiple domains (e.g., device, edge
puting resources. However, taking it a step further, the 55 platform, network). Furthermore, an application slice is
examples described herein show how tenants can be not a monolithic entity. Following the well-known princi-
provided with complete but isolated MEC stacks so that ples of service-oriented architectures, it is built as a set
they can independently orchestrate their assigned re- of atomic data-processing functions which have well de-

11
 21 EP 4 178 169 A1 22

fined functional behaviors and interfaces, referred to plier 1106. The MEC Operator (MOP) 1108 operates and
herein as Application Component Functions (ACFs). An manages ACFs and MEC applications using a MEC Plat-
Application Slice Subnet (APSS) is defined as a set of at form. It is assumed that the MEC Operator 1108 imple-
least one ACF and possible NFs supporting an applica- ments the MEC orchestrator and the ETSI MEC stand-
tion slice. It is essential to point out that application slices 5 ardized interfaces. The MEC Operator 1108 designs,
and network slices are not independent concepts, but builds and operates its MEC platform to offer, together
they are strongly intertwined as application slices de- with the MEC orchestrator, MEC application slice sub-
ployed at the edge of the network require dedicated net- nets to ASPs from ACFs that the ASP has provided as
work slices to deliver the massive amounts of data and an input. The Application Component Function Supplier
network traffic that Application Services typically pro- 10 1106 designs and provides ACFs to ASPs.
duce. As illustrated in detail herein, two distinct models [0083] Interestingly, an ASC 1102 can become an ASP
can be distinguished to allow an application slice to use 1104 while adding new ACFs into consumed ASs and
a network slice. providing new ASs. For instance, assume that ASC C1
[0078] The general concepts that have been described uses from ASP P1 and AS S1 that provides face recog-
can be easily instantiated in the context of MEC, NFV 15 nition capabilities. Then, ASC C1 can integrate into S1
and 5G systems. First, a MEC platform can be used to other ACFs that permit to retrieve a video stream from
offer ASs, which can be deployed as conventional MEC customers’ cameras and ACFs that take control of cus-
applications and implemented using a collection of ACFs. tomers’ door, thus building a new AS S2, which provides
Furthermore, the MEC system can be considered as one automatic door opening based on face recognition. ASC
of the domains over which an application slice can span. 20 C1 then becomes a new ASP P2 selling the previous new
Thus, the MEC management layer should be responsible AS to customers such as house/apartment rental plat-
for the orchestration and deployment of one of the appli- forms. Furthermore, a parallel can be made between FIG.
cation slice subnets that compose the E2E application 11 and FIG. 7. Roles responsible for low layers, namely
slice, referred to as an MEC application slice subnet. A Hardware Supplier 1116, Data Center Service Provider
MEC application slice subnet includes one or more ACFs, 25 1114, NFVI Supplier 1112, and Virtualization Infrastruc-
but it can also include VNFs to support network function- ture Service Provider 1110, remain the same. However,
alities. the Application Component Function Supplier 1106 has
[0079] The VNFs and ACFs have differences and are replaced the VNF (or Equipment) Supplier 708; the MEC
designed for different purposes - network traffic process- Operator 1108 has replaced the Network Operator 706,
ing for VNFs and data processing for the ACFs. Further- 30 the Application Service Provider (ASP) 1104 has re-
more, in a multi-tenant MEC environment, it is likely that placed the CSP 704, and the Application Service Cus-
the MEC applications implementing ACFs may not be tomer (ASC) 1102 has replaced the CSC 702. As shown
orchestrated by the same entity that would orchestrate in FIG. 11, items may operate as a client, a provider, or
principal VNFs. both a client and a provider.
[0080] Nevertheless, there are a lot of common points 35 [0084] A business organization can play one or multi-
in terms of operation and management between a VNF ple operational roles simultaneously. Therefore, without
and an ACF as both rely on the same set of virtualization attempting to be exhaustive in terms of business models,
technologies for their deployment. Thus, both the 3GPP the work as described herein is focused on two categories
management architecture for network slice and the MEC- of organizations whose responsibility boundaries appear
in-NFV architecture can be extended to support the here- 40 to make most of the sense.
in described E2E application slice framework. [0085] The MEC Owner 1105 plays both Hardware
supplier 1116 and NFVI supplier 1112 roles. The fact that
High-level Roles for Application Slice Management the MEC Owner 1105 also manages the virtualization
infrastructure 1110 and the edge servers allows the MEC
[0081] From the APS management (and orchestration) 45 Owner 1105 to dynamically split the infrastructure into
perspective, high-level roles need to be defined in order logical partitions or network slices (i.e. greater degree of
to draw responsibilities boundaries similarly to what was flexibility). The MEC Owner 1105 could be the equivalent
proposed for 5G network slice management (see FIG. of what is referred to as the MEC Host Operator, as the
7). As discussed herein, the focus is on Application Serv- MEC Owner 1105 offers virtualized MEC hosts to its cus-
ices that are offered by a MEC system. 50 tomers. However, the ‘MEC Owner’ 1105 terminology is
[0082] FIG. 11 shows the different roles identified. The preferred to avoid confusion with the ‘MEC Operator’
Application Service Consumer (ASC) 1102 uses Appli- 1108 role.
cation Services. The Application Service Provider (ASP [0086] The MEC Customer 1103: plays both the role
1104) provides Application Services that are built and of the MEC Operator 1108 and ASP 1104. The ASP role
operated upon one or multiple application slice subnets, 55 1104 offers a business interface (with the ASC 1102). In
including MEC application slice subnets. Each of those contrast, the MEC Operator 1108 role offers the actual
application slice subnets is in turn built from a set of ACFs enforcement/implementation (i.e. SLS) of the business
provided by the Application Component Function Sup- objectives (i.e. SLA) agreed across the ASP 1104 busi-

12
 23 EP 4 178 169 A1 24

ness interface. The MEC Operator 1108 role alone can- 1202 includes the data plane (including operating system
not endorse a business organization as it only offers an platform 1230) of the MEC customer 1103/1224. How-
Application Slice (including SLS) and not the Application ever, an alternative overlay model is also possible in
Service (with related SLA). which the E2E network slice 1202 terminates at the net-
[0087] From a business and technical perspective, a 5 work boundary 1302 between the MEC Owner 1222 and
multi-tenancy model in which ASCs 1102 are tenants of the MEC Customer 1224 as per FIG. 13.
a MEC Customer 1103 is preferred, where MEC Cus- [0092] As further shown in FIG. 13, the network and
tomers 1103 are tenants of a MEC Owner 1105. As ex- processing latency 1304 includes that incurred by the
plained further herein, the described multi-tenant MEC operating system platform 1230 of the MEC Server 1214
system supports data isolation (through separated data 10 and the applications 1218 of the MEC server 1214, where
planes) and resource orchestration separation (through the network and processing latency 1304 including the
separate resource orchestrators) between tenants. processing latency 1210 shown in FIG. 12.
[0088] The application service customer 1102 may, as [0093] In the stitching deployment model (shown in
shown in FIG. 11, be an end user, an enterprise, a ver- FIG. 14), it is assumed that the MEC application slice
tical, or other ASP, etc. (collectively 1101). 15 subnet 1402/1224 is a peer of the network slice subnets
(1220, 1222). Virtual appliances using a subnet border
Application-Slice Deployment Models API, such as a virtual gateway or virtual load-balancer,
can be used to interconnect MEC application slice sub-
[0089] Two distinct deployment models are distin- nets 1402 to the adjacent network slice subnet 1222.
guished for the described application slice concept: (i) 20 Such stitching could be a one-to-one interconnection as
the overlay model, and (ii) the stitching model. well as a multiple-to one interconnection. The end-to-end
[0090] The first model assumes that the E2E APS is a application slice 1204 could be seen in this case as the
consumer of a Communication Service offered by the composition of different application slice subnets (UE-
underlying network slices (see FIG. 12). In this case, the operated 1404 or MEC-operated 1402) together with net-
E2E network slice 1202 is responsible for the entire com- 25 work slice subnets (1220, 1222). Latency wise, the MEC
munication latency, including network latency 1208. In application slice subnet 1402 is responsible, in this case,
addition to the E2E network slice 1202, the application for a tiny part 1406 of the network latency budget 1208
slice 1204 caters for processing and possibly storage in addition to the processing and storage latency 1304
latency at both ends, including processing latency 1206 induced by the MEC applications 1218 and their related
at the UE 1212 (by apps 1216 of the UE 1212), the net- 30 ACFs.
work latency 1208, and the processing latency 1210 in- [0094] The different deployment models lead to differ-
curred at the MEC server 1214 end (by apps 1218 of the ent approaches to combine the herein described E2E
MEC server 1214). application slice management/orchestration framework
[0091] The E2E network slice 1202 encompasses not with the 3GPP management architecture, as explained
only the 5G network slice subnet 1220 but also the net- 35 further herein.
work slice subnets (1222, 1224) within the MEC Owner
1222 and MEC Customer 1224 domains. Indeed, as Architecture for Application Slicing in a Multi-tenant MEC
shown in FIG. 12 the herein described slicing framework System
leverages recent advantages on virtualization technolo-
gies that allow a virtualization layer to be composed of 40 [0095] Described herein are new MEC components
multiple nested sub-layers, each using a different virtu- and extensions to the current MEC management archi-
alization paradigm. According to the functional role split tecture needed to support E2E application slicing and
illustrated in FIG. 11, a MEC Owner 1105/1222 (and MEC multi-tenancy within multiple MEC customers. FIG. 15
1226) can use a hypervisor technology 1228 to operate shows another illustrative example of the herein de-
its MEC hosts and to deploy multiple virtualized environ- 45 scribed extended MEC reference architecture 1500. The
ments to MEC customers 1103/1224 (e.g. allocating one primary design rationale of the examples described here-
or more Virtual Machines or VMs (in ETSI NFV terminol- in is that the MEC system should be split into two respon-
ogy, a VM is also designated as ’NFV VNF’) using an sibility domains following a two-layer hierarchical MEC
NFVI). Each virtualized environment includes a full- architecture, where the bottom layer is managed and or-
fledged MEC system used by the MEC Customer 50 chestrated by the MEC Owner, and the top-layer is inde-
1103/1224 to allocate further the resources assigned to pendently managed and orchestrated by MEC Custom-
its VMs to the multiple Application Services it deploys to ers. Such a hierarchical architecture allows a single MEC
its users by applying internal orchestration and manage- deployment to host multiple MEC Customers. Each of
ment criteria. In the case of the MEC customer them has his own MEC network slice subnet (i.e. his ded-
1103/1224, a container-based virtualization technology, 55 icated data plane provided by the MEC Owner) with re-
for instance, could be used as a top virtualization sub- lated management capability. In turn, each MEC Cus-
layer to manage application deployment into its allocated tomer manages and orchestrates his own MEC applica-
virtualized MEC system. In FIG. 12 the E2E network slice tion slices.

13
 25 EP 4 178 169 A1 26

[0096] Implementation-wise, the proposed two-layer form Manager (MECPM) (1540, 1542) to manage the
MEC architecture is enabled by the nested virtualization MEC application slice life-cycle and the MEC Platform
capability of the MEC infrastructure, as anticipated with instance (e.g. embodied as a Docker container). In order
reference to FIG. 12, FIG. 13, and FIG. 14. In the system to stitch the application slice subnets to the adjacent net-
illustrated in FIG. 15, the MEC Owner does not deploy 5 work slice subnets, the MENSM 1514 creates dedicated
following a MEC-in-NFV architecture, but a collection of VNFs (e.g. gateways) (possibly including VNF 1544 and
’ETSI VNFs’ (or VMs) to provide each MEC customer VNF 1546)) that it communicates to the MECO (1534,
with a complete MEC system. The examples described 1536) (or at least the gateway endpoints). A collaboration
herein do not use ’ETSI VNFs’ to deploy MEC applica- between the MECO (1534, 1536) and the MEOO 1502
tions and MEC platforms, but rather to deploy a virtual- 10 could also be necessary in case of MEC application re-
ized MEC environment encompassing virtualized MEC location to enforce new 5GC forwarding rules or tear
hosts and a virtualized MEC management system. Fur- down old ones.
thermore, in the herein described MEC-in-NFV architec- [0098] With regards to the interaction with the 5GC,
ture variant, provided are functional blocks that substitute there are two possible options (1-2):
the MEAO and MEPM-V in the original architecture. Spe- 15
cifically, we substitute the MEAO with the MEC Owner 1. The MEC Owner (e.g. 1502) provides a network
Orchestrator (MEOO) 1502, which is in charge of imple- slice to the MEC Customer (1530, 1532) who directly
menting the policies to select the MEC infrastructures on manages via its MEC Platform (MEP) 5GC forward-
which to deploy a MEC Owner network slice subnet. The ing rules for each application slice (e.g. adds new
MEOO 1502 receives the commands to create, modify 20 DNS rules to the 5GC local DNS servers). This so-
or delete a MEC Owner network slice subnet from a lution allows for better preserving privacy as the MEC
3GPP management function called MEC NSSMF 1504. Customer (1530, 1532) is the only one who deals
Furthermore, the MEOO 1502 collaborates with the MEC with his/her own customers’ UE traffic.
Owner NFVO 1506 to provide a dedicated data plane to
each MEC customer. For the sake of example, it can be 25 2. The MEC Customer MEP (1548, 1550) collabo-
assumed that the MEC Owner offers to each MEC Cus- rates (e.g. via the MECO (1534, 1536) and the ME-
tomer a dedicated Kubernetes cluster, where each K8S OO 1502) with the MEC Owner Network Slice Man-
node is deployed as a ’ETSI VNF’ (or VM) (refer to NFV ager 1514, who ultimately influences 5GC traffic.
VNF 1510 and NFV VNF 1512) in the NFVI 1508, which This solution allows for the MEC Customer (1530,
is connected to the 5G Core (5GC) via a dedicated data 30 1532) to delegate the interaction with 5GC to the
plane (MEC Customer network slice subnet). The second MEC Owner. The MEC Owner can aggregate re-
new functional block is the MEC Network Slice Manager quirements in order to optimize network resources
(MENSM) 1514, which delegates the life-cycle manage- (e.g. bandwidth). Thus, this solution allows for better
ment of the ’ETSI VNFs’ (1510, 1512) to a dedicated network optimization at the MEC Owner infrastruc-
VNFM (one or more of VNFM 1516-1, VNFM 1516-2, 35 ture but does not preserve privacy. Also, it may be
and/or VNFM 1516-3), while the MEC Network Slice less scalable as the number of UEs increases.
Manager (MENSM) 1514 is responsible for the manage-
ment of the network slice subnet (data plane) parame- [0099] The MEC Reference Architecture [ETSI, Multi-
ters. For instance, the MEC Network Slice Manager access edge computing (mec); framework and reference
(MENSM) 1514 can reserve network bandwidth between 40 architecture, Group Specification ETSI GS MEC 003
MEC hosts for a given MEC Customer (1530, 1532). V2.2.1, ETSI ISG MEC (12 2020)] shows a single MEC
Moreover, the MEC Network Slice Manager 1514 could orchestrator controlling a single virtualization infrastruc-
behave like an 3GPP Application Function (AF) which ture and managing the instantiation of all MEC applica-
interacts with the 5GC to synchronize data plane forward- tions. The herein described MEC architecture implies a
ing rules to realize local breakout traffic to/from MEC ap- 45 split of the MEC orchestrator into a MEC Customer Or-
plications (refer to MEC app 1518, MEC app 1520, MEC chestrator (MECO) (1534, 1536) and a MEC Owner Or-
app 1522, MEC app 1524, MEC app 1526, MEC app chestrator (MEOO) 1502. While the MECO (1534, 1536)
1528). is responsible for the MEC Platform, MEC applications
[0097] As initially described, each MEC Customer (1518, 1520, 1522, 1524, 1526, 1528), MEC application
(1530, 1532) manages and orchestrates its own MEC 50 slices (1552, 1554, 1556, 1558, 1560, 1562), and related
application slices within the assigned virtualized MEC external interfaces (line couplings between entities
system 1500. To this end, each MEC Customer (1530, shown in FIG. 15, including from MEC customer 1530
1532) implements a MEC Customer Orchestrator and MEC customer 1532), the MEOO 1502 is responsi-
(MECO) (1534, 1536), which receives the commands to ble for the hardware, the NFVI 1508, MEC NFVI slices
create, modify or delete MEC application slices from a 55 (relating perhaps to VNF 1510 and VNF 1512) (especially
management function called MEC APSSMF 1538 (refer MEC network slices) and related external interfaces (line
also to FIG. 5 and FIG. 16). Furthermore, the MECO couplings between entities shown in FIG. 15, including
(1534, 1536) collaborates with the MEC Customer Plat- from MEOO 1502).

14
 27 EP 4 178 169 A1 28

[0100] FIG. 15 shows a legend including MEC APSs (non-expert user) does not need to perform the afore-
in MEC customer #1 (1564) and MEC APSs in MEC cus- mentioned E2E AS SLA dichotomy. In this stitching mod-
tomer #2 (1566). The MEC APSs in MEC customer #1 el, the APSMF 506 is responsible to communicate direct-
1530 include MEC APSs 1552, 1554, and 1556, while ly with the NSMF 510 to manage the network slice subnet
the MEC APSs in customer #2 1532 include MEC APSs 5 associated with the APS.
1558, 1560,and 1562. There is a further split within MEC [0105] The management of the network slice via the
customer of ASP #1 1530, wherein MEC appl 1518 and NSMF 510 and the one of per-domain network slice sub-
MEC app2 1520 belong to MEC ASC #1, while MEC appl nets via NSSMFs are well-defined by 3GPP, and they do
1522 belongs to MEC ASC #2. MEC appl 1524, MEC not need to be extended.
app2 1526, and MEC app3 1528 belong to MEC ASC #3 10 [0106] FIG. 5 further distinguishes between business
of MEC Customer of ASP #2 1532. User app LCM proxy interfaces (collectively 534) and internal interfaces (col-
1568 is associated with MEC Customer of ASP #1 1530, lectively 536).
and user app LCM proxy 1570 is associated with MEC [0107] The experience in implementing the manage-
Customer of ASP #2 1532. ment architecture provided by the exampled herein is
15 also described. As shown in FIG. 16, several orchestra-
Application Slice Management Architecture tion entities are involved in the management of the var-
ious application slice subnets. The focus is on implement-
[0101] With the aforementioned new roles and archi- ing the MEC Customer Orchestrator 1602 using a pop-
tecture in mind, the 3GPP network slice management ular open-source container orchestration platform. Fur-
architecture could also be augmented to manage and 20 ther described in detail are the interfaces and data mod-
orchestrate application slices as illustrated in FIG. 5, por- els needed to interact with the MEC APSSMF 1604, al-
tions of the description of which are herein repeated, for lowing the deployment of isolated MEC application slice
further clarity. It is assumed that an ASC relies on a web subnets composed of ACFs in the form of Docker con-
portal 524 to request an application service with a given tainers (1606-1, 1606-2, 1606-3, 1606-4).
SLA from a catalogue of offered ASs (see FIG. 18 and 25 [0108] Also shown in FIG. 16 is OS-based VI 1608
corresponding description for more details on how to im- managed with MEO-C 1602, Core NSSMF 1610 which
plement such service catalogue 1802 of CFS portal interfaces with NFVO 1614, and MES NSSMF 1612
1803). The business interaction of the web portal 524 which interfaces with MEO-O 1616. The NFVO 1614 in-
can happen in two manners depending on the AS/APS terfaces with the MEO-O 1616 which interfaces with the
deployment models that is used in the system (refer to 30 MEO-C 1602. The NFVO 1614 and MEO-O 1616 both
FIG. 12, FIG. 13, and FIG. 14 and corresponding descrip- instantiate/manage the NFVI 1618. FIG. 16 also shows
tions). In both cases, the web portal 524 communicates several virtual machines (VM) (1620-1, 1620-2, 1620-3,
with a new management function, called Application 1620-4, and 1620-5) .
Service Management Function (ASMF) 504, which is re- [0109] FIG. 17 is an example apparatus 1700, which
sponsible for translating the SLA of the requested AS to 35 may be implemented in hardware, configured to imple-
the specification of an APS and to trigger the creation of ment the examples described herein. The apparatus
the APS instance by contacting a new management func- 1700 comprises at least one processor 1702 (an FPGA
tion called Application Slice Management Function and/or CPU), at least one memory 1704 including com-
(APSMF) 506. The APSMF 506 splits the APS into mul- puter program code 1705, wherein at least one memory
tiple subnets, one for each domain over which the re- 40 1704 and the computer program code 1705 are config-
quested APS spans, including possibly the network and ured to, with at least one processor 1702, cause the ap-
the AS’s endpoints, namely the UE 110 requesting the paratus 1700 to implement circuitry, a process, compo-
AS, and the edge system instantiating the AS. To this nent, module, or function (collectively control 1706) to
end, introduced herein is the new Application Slice Sub- implement the examples described herein, including mul-
net Management Function (APSSMF) (512, 522), which 45 ti-access edge computing slicing. The memory 1704 may
applies the APSS life-cycle management commands be a non-transitory memory, a transitory memory, a vol-
within the two potential domains that are relevant for an atile memory, or a non-volatile memory.
application slice subnet, namely the UE 512 and the MEC [0110] The apparatus 1700 optionally includes a dis-
522. play and/or I/O interface 1708 that may be used to display
[0102] In the overlay model (label 1 501), the E2E AS- 50 aspects or a status of the methods described herein (e.g.,
MF 504 is also responsible for translating the E2E AS as one of the methods is being performed or at a subse-
SLA into E2E CS SLA and for requiring the adapted E2E quent time), or to receive input from a user such as with
CS from the CSMF 508. using a keypad. The apparatus 1700 includes one or
[0103] In one of the stitching models (label 2 502), the more network (N/W) interfaces (I/F(s)) 1710. The N/W
web portal 524 (expert user) is also responsible for break- 55 I/F(s) 1710 may be wired and/or wireless and communi-
ing the E2E AS SLA into AS SLA and E2E CS SLA and cate over the Internet/other network(s) via any commu-
for requiring the adapted E2E CS from the CSMF 508. nication technique. The N/W I/F(s) 1710 may comprise
[0104] Alternatively (label 3 503), the web portal 524 one or more transmitters and one or more receivers. The

15
 29 EP 4 178 169 A1 30

N/W I/F(s) 1710 may comprise standard well-known synchronizing, using the interface, multi-access edge
components such as an amplifier, filter, frequency-con- computing forwarding rules. Method 1800 may be per-
verter, (de)modulator, and encoder/decoder circuitries formed with an MEC-OO (e.g. 402).
and one or more antennas. [0114] FIG. 19 is an example method 1900 to imple-
[0111] The apparatus 1700 to implement the function- 5 ment the example embodiments described herein. At
ality of control 1706 may be UE 110, RAN node 170, or 1902, the method includes managing first resources with-
network element(s) 190. Thus, processor 1702 may cor- in a network slice subnet of a multi-access edge comput-
respond respectively to processor(s) 120, processor(s) ing customer, the first resources managed within a multi-
152 and/or processor(s) 175, memory 1704 may corre- access edge computing customer orchestrator. At 1904,
spond respectively to memory(ies) 125, memory(ies) 155 10 the method includes establishing an interface with a mul-
and/or memory(ies) 171, computer program code 1705 ti-access edge computing owner orchestrator. At 1906,
may correspond respectively to computer program code the method includes synchronizing, using the interface,
123, module 140-1, module 140-2, and/or computer pro- the first resources with second resources, the second
gram code 153, module 150-1, module 150-2, and/or resources managed within the network slice subnet of
computer program code 173, and N/W I/F(s) 1710 may 15 the multi-access edge computing customer, the second
correspond respectively to N/W I/F(s) 161 and/or N/W resources managed within the multi-access edge com-
I/F(s) 180. Alternatively, apparatus 1700 may not corre- puting owner orchestrator. At 1908, the method includes
spond to either of UE 110, RAN node 170, or network partitioning the first resources into smaller application
element (s) 190, as apparatus 1700 may be part of a self- slices. At 1910, the method includes synchronizing, using
organizing/optimizing network (SON) node, such as in a 20 the interface, multi-access edge computing forwarding
cloud. Apparatus 1700 may correspond to any of the ap- rules. Method 1900 may be performed with an MEC-CO
paratuses shown in the other figures such as the MEC (e.g. 404).
owner orchestrator 402, the MEC customer orchestrator [0115] FIG. 20 is an example method 2000 to imple-
404, the ASMF 504, the UE 110, the web portal 524, the ment the example embodiments described herein. At
APSMF 506, or the APSSMF 430. The apparatus 1700 25 2002, the method includes receiving, from a manage-
may also be distributed throughout the network 100 in- ment web portal, a request for an application service. At
cluding within and between apparatus 1700 and any net- 2004, the method includes translating, with an application
work element (such as a network control element (NCE) service management function, a service level agreement
190 and/or the RAN node 170 and/or the UE 110). of the requested application service to a specification of
[0112] Interface 1712 enables data communication be- 30 an application slice. At 2006, the method includes trans-
tween the various items of apparatus 1700, as shown in lating the service level agreement of the application serv-
FIG. 17. For example, the interface 1712 may be one or ice into a service level agreement of a communication
more buses such as address, data, or control buses, and service. At 2008, the method includes establishing an
may include any interconnection mechanism, such as a interface with a communication service management
series of lines on a motherboard or integrated circuit, 35 function for communication of the service level agree-
fiber optics or other optical communication equipment, ment of the communication service. Method 2000 may
and the like. Computer program code 1705, including be performed with an ASMF (e.g. 504).
control 1706 may comprise object-oriented software con- [0116] FIG. 21 is an example method 2100 to imple-
figured to pass data/messages between objects within ment the example embodiments described herein. At
computer program code 1705. The apparatus 1700 need 40 2102, the method includes managing an application serv-
not comprise each of the features mentioned, or may ice using a management web portal. At 2104, the method
comprise other features as well. includes requesting, using the management web portal,
[0113] FIG. 18 is an example method 1800 to imple- the application service with a given service level agree-
ment the example embodiments described herein. At ment from a catalogue of offered application services. At
1802, the method includes managing first resources with- 45 2106, the method includes communicating, using the
in a network slice subnet of a multi-access edge comput- management web portal, with an application service
ing customer, the first resources managed within a multi- management function. At 2108, the method includes
access edge computing owner orchestrator. At 1804, the wherein the application service management function is
method includes establishing an interface with a multi- configured to translate the service level agreement of the
access edge computing customer orchestrator. At 1806, 50 requested application service to a specification of an ap-
the method includes synchronizing, using the interface, plication slice, and to trigger a creation of an application
the first resources with second resources, the second slice instance with contacting an application slice man-
resources managed within the network slice subnet of agement function. Method 2100 may be performed with
the multi-access edge computing customer, the second a UE (e.g. 110) or web portal (e.g. 524) .
resources managed within the multi-access edge com- 55 [0117] FIG. 22 is an example method 2200 to imple-
puting customer orchestrator. At 1808, the method in- ment the example embodiments described herein. At
cludes wherein the second resources are partitioned into 2202, the method includes establishing an interface with
smaller application slices. At 1810, the method includes an application service management function. At 2204,

16
 31 EP 4 178 169 A1 32

the method includes receiving, over the interface from or applications processor integrated circuit for a mobile
an application service management function with an ap- phone or a similar integrated circuit in a server, a cellular
plication slice management function, a request to trigger network device, or another network device.
a creation of an application slice. At 2206, the method [0122] The following examples (1-54) are provided and
includes splitting the application slice into multiple sub- 5 described herein.
nets, one subnet of the multiple subnets corresponding [0123] Example 1: An example apparatus includes at
to a domain over which the application slice spans. Meth- least one processor; and at least one memory including
od 2200 may be performed with an APSMF (e.g. 506). computer program code; wherein the at least one mem-
[0118] FIG. 23 is an example method 2300 to imple- ory and the computer program code are configured to,
ment the example embodiments described herein. At 10 with the at least one processor, cause the apparatus at
2302, the method includes establishing an interface with least to: manage first resources within a network slice
an application slice management function. At 2304, the subnet of a multi-access edge computing customer, the
method includes applying an application slice subnet li- first resources managed within a multi-access edge com-
fecycle management command with a domain relevant puting owner orchestrator; establish an interface with a
for a user equipment slice subnet, and with a domain 15 multi-access edge computing customer orchestrator;
relevant for a multi-access edge computing subnet. synchronize, using the interface, the first resources with
Method 2300 may be performed with an APSSMF (e.g. second resources, the second resources managed with-
430). in the network slice subnet of the multi-access edge com-
[0119] References to a ’computer’, ’processor’, etc. puting customer, the second resources managed within
should be understood to encompass not only computers 20 the multi-access edge computing customer orchestrator;
having different architectures such as single/multi-proc- wherein the second resources are partitioned into smaller
essor architectures and sequential or parallel architec- application slices; and synchronize, using the interface,
tures but also specialized circuits such as field-program- multi-access edge computing forwarding rules.
mable gate arrays (FPGAs), application specific circuits [0124] Example 2: The apparatus of example 1, where-
(ASICs), signal processing devices and other processing 25 in the network slice subnet of the multi-access edge com-
circuitry. References to computer program, instructions, puting customer is managed with a network slice subnet
code etc. should be understood to encompass software management function of the multi-access edge comput-
for a programmable processor or firmware such as, for ing customer.
example, the programmable content of a hardware de- [0125] Example 3: The apparatus of any one of exam-
vice whether instructions for a processor, or configuration 30 ples 1 to 2, wherein an application slice of the smaller
settings for a fixed-function device, gate array or pro- application slices is managed with an application slice
grammable logic device etc. management function of the multi-access edge comput-
[0120] The memory(ies) as described herein may be ing customer.
implemented using any suitable data storage technology, [0126] Example 4: The apparatus of example 3, where-
such as semiconductor based memory devices, flash 35 in the application slice management function is config-
memory, magnetic memory devices and systems, optical ured to delegate lifecycle management of constituent ap-
memory devices and systems, non-transitory memory, plication slice subnets to a different domain application
transitory memory, fixed memory and removable mem- slice subnet management function.
ory. The memory(ies) may comprise a database for stor- [0127] Example 5: The apparatus of example 4, where-
ing data. 40 in the application slice subnet management function
[0121] As used herein, the term ’circuitry’ may refer to manages an application component function.
the following: (a) hardware circuit implementations, such [0128] Example 6: The apparatus of any one of exam-
as implementations in analog and/or digital circuitry, and ples 1 to 5, wherein the at least one memory and the
(b) combinations of circuits and software (and/or computer program code are further configured to, with
firmware), such as (as applicable): (i) a combination of 45 the at least one processor, cause the apparatus at least
processor(s) or (ii) portions of processor(s)/software in- to: receive an order from a multi-access edge computing
cluding digital signal processor(s), software, and mem- network slice subnet management function to create,
ory(ies) that work together to cause an apparatus to per- modify, or delete a multi-access edge computing network
form various functions, and (c) circuits, such as a micro- slice subnet.
processor(s) or a portion of a microprocessor(s), that re- 50 [0129] Example 7: The apparatus of example 6, where-
quire software or firmware for operation, even if the soft- in the at least one memory and the computer program
ware or firmware is not physically present. As a further code are further configured to, with the at least one proc-
example, as used herein, the term ’circuitry’ would also essor, cause the apparatus at least to: collaborate with
cover an implementation of merely a processor (or mul- a multi-access edge computing owner virtual environ-
tiple processors) or a portion of a processor and its (or 55 ment orchestrator to manage a lifecycle of the multi-ac-
their) accompanying software and/or firmware. The term cess edge computing network slice subnet.
’circuitry’ would also cover, for example and if applicable [0130] Example 8: The apparatus of any one of exam-
to the particular element, a baseband integrated circuit ples 1 to 7, wherein the at least one memory and the

17
 33 EP 4 178 169 A1 34

computer program code are further configured to, with customer, the second resources managed within the mul-
the at least one processor, cause the apparatus at least ti-access edge computing owner orchestrator; partition
to: collaborate with a multi-access edge computing net- the first resources into smaller application slices; and
work slice manager to manage network slice subnet data synchronize, using the interface, multi-access edge com-
plane parameters. 5 puting forwarding rules.
[0131] Example 9: The apparatus of example 8, where- [0136] Example 14: The apparatus of example 13,
in the at least one memory and the computer program wherein the network slice subnet of the multi-access
code are further configured to, with the at least one proc- edge computing customer is managed with a network
essor, cause the apparatus at least to: establish an in- slice subnet management function of the multi-access
terface between the multi-access edge computing net- 10 edge computing customer.
work slice subnet manager and a multi-access edge com- [0137] Example 15: The apparatus of any one of ex-
puting customer platform manager associated with the amples 13 to 14, wherein an application slice of the small-
multi-access edge computing customer orchestrator. er application slices is managed with an application slice
[0132] Example 10: The apparatus of any one of ex- management function of the multi-access edge comput-
amples 8 to 9, wherein the at least one memory and the 15 ing customer.
computer program code are further configured to, with [0138] Example 16: The apparatus of example 15,
the at least one processor, cause the apparatus at least wherein the application slice management function is
to: act as 3GPP Application Function (AF) and influence configured to delegate lifecycle management of constit-
3GPP (such as, for example, 5G or 6G) Core traffics uent application slice subnets to a different domain ap-
from/towards application slices of the multi-access edge 20 plication slice subnet management function.
computing customer. [0139] Example 17: The apparatus of example 16,
[0133] Example 11: The apparatus of any one of ex- wherein the application slice subnet management func-
amples 8 to 10, wherein the at least one memory and the tion manages an application component function.
computer program code are further configured to, with [0140] Example 18: The apparatus of any one of ex-
the at least one processor, cause the apparatus at least 25 amples 13 to 17, wherein the at least one memory and
to: synchronize, using an interface of a multi-access edge the computer program code are further configured to,
computing customer platform manager associated with with the at least one processor, cause the apparatus at
the multi-access edge computing customer orchestrator, least to: receive an order from a multi-access edge com-
first 5G forwarding rules with second 5G forwarding rules; puting application slice subnet management function to
wherein the first 5G forwarding rules are associated with 30 create, modify, or delete a multi-access edge computing
the collaboration with the multi-access edge computing application slice of the smaller application slices.
network subnet slice manager to manage network slice [0141] Example 19: The apparatus of example 18,
subnet data plane parameters, and the second 5G for- wherein the at least one memory and the computer pro-
warding rules are associated with the multi-access edge gram code are further configured to, with the at least one
computing customer platform manager associated with 35 processor, cause the apparatus at least to: collaborate
the multi-access edge computing customer orchestrator. with a multi-access edge computing customer virtual en-
[0134] Example 12: The apparatus of any one of ex- vironment orchestrator to manage a lifecycle of the multi-
amples 1 to 11, wherein the at least one memory and the access edge computing application slice.
computer program code are further configured to, with [0142] Example 20: The apparatus of any one of ex-
the at least one processor, cause the apparatus at least 40 amples 13 to 19, wherein the at least one memory and
to: create a gateway to stitch a multi-access edge com- the computer program code are further configured to,
puting application slice data plane to a multi-access edge with the at least one processor, cause the apparatus at
computing network slice subnet data plane; and commu- least to: collaborate with a multi-access edge computing
nicate the gateway to the multi-access edge computing customer platform manager to manage a multi-access
customer orchestrator. 45 edge computing platform instance.
[0135] Example 13: An apparatus includes at least one [0143] Example 21: The apparatus of example 20,
processor; and at least one memory including computer wherein the at least one memory and the computer pro-
program code; wherein the at least one memory and the gram code are further configured to, with the at least one
computer program code are configured to, with the at processor, cause the apparatus at least to: act as 3GPP
least one processor, cause the apparatus at least to: 50 Application Function (AF) and influence 3GPP (such as,
manage first resources within a network slice subnet of for example, 5G or 6G) Core traffics from/towards appli-
a multi-access edge computing customer, the first re- cation slices of the multi-access edge computing cus-
sources managed within a multi-access edge computing tomer.
customer orchestrator; establish an interface with a multi- [0144] Example 22: The apparatus of any one of ex-
access edge computing owner orchestrator; synchro- 55 amples 20 to 21, wherein the at least one memory and
nize, using the interface, the first resources with second the computer program code are further configured to,
resources, the second resources managed within the with the at least one processor, cause the apparatus at
network slice subnet of the multi-access edge computing least to: establish an interface between the multi-access

18
 35 EP 4 178 169 A1 36

edge computing customer platform manager and a multi- tion, the second service level agreement dedicated to
access edge computing network slice subnet manager application service endpoints, the application service
associated with the multi-access edge computing owner endpoints comprising a user equipment and a multi-ac-
orchestrator. cess edge computing device.
[0145] Example 23: The apparatus of example 22, 5 [0150] Example 28: An apparatus includes at least one
wherein the at least one memory and the computer pro- processor; and at least one memory including computer
gram code are further configured to, with the at least one program code; wherein the at least one memory and the
processor, cause the apparatus at least to: synchronize, computer program code are configured to, with the at
using the interface between the multi-access edge com- least one processor, cause the apparatus at least to:
puting customer platform manager and the multi-access 10 manage an application service using a management web
edge computing network slice manager associated with portal; request, using the management web portal, the
the multi-access edge computing owner orchestrator, application service with a given service level agreement
first 5G forwarding rules with second 5G forwarding rules; from a catalogue of offered application services; and
wherein the first 5G forwarding rules are associated with communicate, using the management web portal, with
a collaboration with a multi-access edge computing net- 15 an application service management function; wherein the
work slice manager to manage network slice data plane application service management function is configured
parameters, and the second 5G forwarding rules are as- to translate the service level agreement of the requested
sociated with a collaboration with a multi-access edge application service to a specification of an application
computing customer platform manager to manage a mul- slice, and to trigger a creation of an application slice in-
ti-access edge computing platform instance. 20 stance with contacting an application slice management
[0146] Example 24: The apparatus of any one of ex- function.
amples 13 to 23, wherein the at least one memory and [0151] Example 29: The apparatus of example 28,
the computer program code are further configured to, wherein the at least one memory and the computer pro-
with the at least one processor, cause the apparatus at gram code are further configured to, with the at least one
least to: receive a communication of a gateway from the 25 processor, cause the apparatus at least to: manage at
multi-access edge computing owner orchestrator; least one network service, the network service being re-
wherein the gateway is configured to stitch a multi-access lated to the application service.
edge computing application slice data plane to a multi- [0152] Example 30: The apparatus of any one of ex-
access edge computing network slice data plane. amples 28 to 29, wherein: the application slice instance
[0147] Example 25: An apparatus includes at least one 30 is split with the application slice management function
processor; and at least one memory including computer into multiple subnets, one subnet of the multiple subnets
program code; wherein the at least one memory and the corresponding to a domain over which the application
computer program code are configured to, with the at slice spans; and an application slice subnet management
least one processor, cause the apparatus at least to: re- function applies an application slice subnet lifecycle man-
ceive, from a management web portal, a request for an 35 agement command with a domain relevant for a user
application service; translate, with an application service equipment slice subnet, and with a domain relevant for
management function, a service level agreement of the a multi-access edge computing subnet.
requested application service to a specification of an ap- [0153] Example 31: The apparatus of any one of ex-
plication slice; translate the service level agreement of amples 28 to 30, wherein the at least one memory and
the application service into a service level agreement of 40 the computer program code are further configured to,
a communication service; and establish an interface with with the at least one processor, cause the apparatus at
a communication service management function for com- least to: partition the service level agreement in into a
munication of the service level agreement of the commu- first service level agreement and a second service level
nication service. agreement, the first service level agreement being inde-
[0148] Example 26: The apparatus of example 25, 45 pendent of the second service level agreement.
wherein the at least one memory and the computer pro- [0154] Example 32: An apparatus includes at least one
gram code are further configured to, with the at least one processor; and at least one memory including computer
processor, cause the apparatus at least to: establish an program code; wherein the at least one memory and the
interface with an application slice management function. computer program code are configured to, with the at
[0149] Example 27: The apparatus of any one of ex- 50 least one processor, cause the apparatus at least to: es-
amples 25 to 26, wherein the management web portal is tablish an interface with an application service manage-
configured to partition the service level agreement into a ment function; receive, over the interface from an appli-
first service level agreement and a second service level cation service management function with an application
agreement, the first service level agreement being inde- slice management function, a request to trigger a creation
pendent of the second service level agreement, the first 55 of an application slice; and split the application slice into
service level agreement related to a communication serv- multiple subnets, one subnet of the multiple subnets cor-
ice that the application service management function re- responding to a domain over which the application slice
quests from a communication service management func- spans.

19
 37 EP 4 178 169 A1 38

[0155] Example 33: The apparatus of example 32, sources into smaller application slices; and synchroniz-
wherein the at least one memory and the computer pro- ing, using the interface, multi-access edge computing for-
gram code are further configured to, with the at least one warding rules.
processor, cause the apparatus at least to: split the ap- [0161] Example 39: A method includes receiving, from
plication slice into multiple subnets, one subnet of the 5 a management web portal, a request for an application
multiple subnets corresponding to a domain over which service; translating, with an application service manage-
the application slice spans. ment function, a service level agreement of the requested
[0156] Example 34: The apparatus of any one of ex- application service to a specification of an application
amples 32 to 33, wherein the at least one memory and slice; translating the service level agreement of the ap-
the computer program code are further configured to, 10 plication service into a service level agreement of a com-
with the at least one processor, cause the apparatus at munication service; and establishing an interface with a
least to: establish an interface with an application slice communication service management function for com-
subnet management function. munication of the service level agreement of the commu-
[0157] Example 35: The apparatus of any one of ex- nication service.
amples 32 to 34, wherein the at least one memory and 15 [0162] Example 40: A method includes managing an
the computer program code are further configured to, application service using a management web portal; re-
with the at least one processor, cause the apparatus at questing, using the management web portal, the appli-
least to: establish a second interface with a network slice cation service with a given service level agreement from
management function; wherein the application slice man- a catalogue of offered application services; and commu-
agement function is configured to communicate directly 20 nicating, using the management web portal, with an ap-
with the network slice management function over the sec- plication service management function; wherein the ap-
ond interface to manage network slice subnets associ- plication service management function is configured to
ated with the application slice. translate the service level agreement of the requested
[0158] Example 36: An apparatus includes at least one application service to a specification of an application
processor; and at least one memory including computer 25 slice, and to trigger a creation of an application slice in-
program code; wherein the at least one memory and the stance with contacting an application slice management
computer program code are configured to, with the at function.
least one processor, cause the apparatus at least to: es- [0163] Example 41: A method includes establishing an
tablish an interface with an application slice management interface with an application service management func-
function; and apply an application slice subnet lifecycle 30 tion; receiving, over the interface from an application
management command with a domain relevant for a user service management function with an application slice
equipment slice subnet, and with a domain relevant for management function, a request to trigger a creation of
a multi-access edge computing subnet. an application slice; and splitting the application slice into
[0159] Example 37: A method includes managing first multiple subnets, one subnet of the multiple subnets cor-
resources within a network slice subnet of a multi-access 35 responding to a domain over which the application slice
edge computing customer, the first resources managed spans.
within a multi-access edge computing owner orchestra- [0164] Example 42: A method includes establishing an
tor; establishing an interface with a multi-access edge interface with an application slice management function;
computing customer orchestrator; synchronizing, using and applying an application slice subnet lifecycle man-
the interface, the first resources with second resources, 40 agement command with a domain relevant for a user
the second resources managed within the network slice equipment slice subnet, and with a domain relevant for
subnet of the multi-access edge computing customer, a multi-access edge computing subnet.
the second resources managed within the multi-access [0165] Example 43: A non-transitory program storage
edge computing customer orchestrator; wherein the sec- device readable by a machine, tangibly embodying a pro-
ond resources are partitioned into smaller application 45 gram of instructions executable with the machine for per-
slices; and synchronizing, using the interface, multi-ac- forming operations, the operations comprising: manag-
cess edge computing forwarding rules. ing first resources within a network slice subnet of a multi-
[0160] Example 38: A method includes managing first access edge computing customer, the first resources
resources within a network slice subnet of a multi-access managed within a multi-access edge computing owner
edge computing customer, the first resources managed 50 orchestrator; establishing an interface with a multi-ac-
within a multi-access edge computing customer orches- cess edge computing customer orchestrator; synchro-
trator; establishing an interface with a multi-access edge nizing, using the interface, the first resources with second
computing owner orchestrator; synchronizing, using the resources, the second resources managed within the
interface, the first resources with second resources, the network slice subnet of the multi-access edge computing
second resources managed within the network slice sub- 55 customer, the second resources managed within the mul-
net of the multi-access edge computing customer, the ti-access edge computing customer orchestrator; where-
second resources managed within the multi-access edge in the second resources are partitioned into smaller ap-
computing owner orchestrator; partitioning the first re- plication slices; and synchronizing, using the interface,

20
 39 EP 4 178 169 A1 40

multi-access edge computing forwarding rules. responding to a domain over which the application slice
[0166] Example 44: A non-transitory program storage spans.
device readable by a machine, tangibly embodying a pro- [0170] Example 48: A non-transitory program storage
gram of instructions executable with the machine for per- device readable by a machine, tangibly embodying a pro-
forming operations, the operations comprising: manag- 5 gram of instructions executable with the machine for per-
ing first resources within a network slice subnet of a multi- forming operations, the operations comprising: establish-
access edge computing customer, the first resources ing an interface with an application slice management
managed within a multi-access edge computing custom- function; and applying an application slice subnet lifecy-
er orchestrator; establishing an interface with a multi-ac- cle management command with a domain relevant for a
cess edge computing owner orchestrator; synchronizing, 10 user equipment slice subnet, and with a domain relevant
using the interface, the first resources with second re- for a multi-access edge computing subnet.
sources, the second resources managed within the net- [0171] Example 49: An apparatus includes means for
work slice subnet of the multi-access edge computing managing first resources within a network slice subnet
customer, the second resources managed within the mul- of a multi-access edge computing customer, the first re-
ti-access edge computing owner orchestrator; partition- 15 sources managed within a multi-access edge computing
ing the first resources into smaller application slices; and owner orchestrator; means for establishing an interface
synchronizing, using the interface, multi-access edge with a multi-access edge computing customer orchestra-
computing forwarding rules. tor; means for synchronizing, using the interface, the first
[0167] Example 45: A non-transitory program storage resources with second resources, the second resources
device readable by a machine, tangibly embodying a pro- 20 managed within the network slice subnet of the multi-
gram of instructions executable with the machine for per- access edge computing customer, the second resources
forming operations, the operations comprising: receiving, managed within the multi-access edge computing cus-
from a management web portal, a request for an appli- tomer orchestrator; wherein the second resources are
cation service; translating, with an application service partitioned into smaller application slices; and means for
management function, a service level agreement of the 25 synchronizing, using the interface, multi-access edge
requested application service to a specification of an ap- computing forwarding rules.
plication slice; translating the service level agreement of [0172] Example 50: An apparatus includes means for
the application service into a service level agreement of managing first resources within a network slice subnet
a communication service; and establishing an interface of a multi-access edge computing customer, the first re-
with a communication service management function for 30 sources managed within a multi-access edge computing
communication of the service level agreement of the customer orchestrator; means for establishing an inter-
communication service. face with a multi-access edge computing owner orches-
[0168] Example 46: A non-transitory program storage trator; means for synchronizing, using the interface, the
device readable by a machine, tangibly embodying a pro- first resources with second resources, the second re-
gram of instructions executable with the machine for per- 35 sources managed within the network slice subnet of the
forming operations, the operations comprising: manag- multi-access edge computing customer, the second re-
ing an application service using a management web por- sources managed within the multi-access edge comput-
tal; requesting, using the management web portal, the ing owner orchestrator; means for partitioning the first
application service with a given service level agreement resources into smaller application slices; and means for
from a catalogue of offered application services; and 40 synchronizing, using the interface, multi-access edge
communicating, using the management web portal, with computing forwarding rules.
an application service management function; wherein the [0173] Example 51: An apparatus includes means for
application service management function is configured receiving, from a management web portal, a request for
to translate the service level agreement of the requested an application service; means for translating, with an ap-
application service to a specification of an application 45 plication service management function, a service level
slice, and to trigger a creation of an application slice in- agreement of the requested application service to a spec-
stance with contacting an application slice management ification of an application slice; means for translating the
function. service level agreement of the application service into a
[0169] Example 47: A non-transitory program storage service level agreement of a communication service; and
device readable by a machine, tangibly embodying a pro- 50 means for establishing an interface with a communication
gram of instructions executable with the machine for per- service management function for communication of the
forming operations, the operations comprising: establish- service level agreement of the communication service.
ing an interface with an application service management [0174] Example 52: An apparatus includes means for
function; receiving, over the interface from an application managing an application service using a management
service management function with an application slice 55 web portal; means for requesting, using the management
management function, a request to trigger a creation of web portal, the application service with a given service
an application slice; and splitting the application slice into level agreement from a catalogue of offered application
multiple subnets, one subnet of the multiple subnets cor- services; and means for communicating, using the man-

21
 41 EP 4 178 169 A1 42

agement web portal, with an application service manage- tion

ment function; wherein the application service manage- ASP application service provider
ment function is configured to translate the service level CFS customer facing service
agreement of the requested application service to a spec- CS communication service
ification of an application slice, and to trigger a creation 5 CSMF communication service management
of an application slice instance with contacting an appli- function
cation slice management function. CN core network
[0175] Example 53: An apparatus includes means for CPU central processing unit
establishing an interface with an application service man- CU central unit or centralized unit
agement function; means for receiving, over the interface 10 DC dual connectivity
from an application service management function with D-plane data plane
an application slice management function, a request to DSP digital signal processor
trigger a creation of an application slice; and means for DU distributed unit
splitting the application slice into multiple subnets, one E2E end to end
subnet of the multiple subnets corresponding to a domain 15 EP AS SLA endpoint-related application service
over which the application slice spans. SLA
[0176] Example 54: An apparatus includes means for eNB evolved Node B (e.g., an LTE base sta-
establishing an interface with an application slice man- tion)
agement function; and means for applying an application EN-DC E-UTRA-NR dual connectivity
slice subnet lifecycle management command with a do- 20 en-gNB node providing NR user plane and con-
main relevant for a user equipment slice subnet, and with trol plane protocol terminations towards
a domain relevant for a multi-access edge computing the UE, and acting as a secondary node
subnet. in EN-DC
[0177] It should be understood that the foregoing de- etcd "/etc distributed", distributed key-value
scription is only illustrative. Various alternatives and mod- 25 store
ifications may be devised by those skilled in the art. For ETSI European Telecommunications Stand-
example, features recited in the various dependent ards Institute
claims could be combined with each other in any suitable E-UTRA evolved universal terrestrial radio ac-
combination(s). In addition, features from different em- cess, i.e., the LTE radio access technol-
bodiments described above could be selectively com- 30 ogy
bined into a new embodiment. Accordingly, this descrip- F1 interface between the CU and the DU
tion is intended to embrace all such alternatives, modifi- FPGA field-programmable gate array
cations and variances which fall within the scope of the gNB base station for 5G/NR, i.e., a node pro-
appended claims. viding NR user plane and control plane
[0178] The following acronyms and abbreviations that 35 protocol terminations towards the UE,
may be found in the specification and/or the drawing fig- and connected via the NG interface to
ures are defined as follows (the abbreviations and acro- the 5GC
nyms may be appended with each other or with other ID identifier
characters using e.g. a dash or hyphen): I/F interface
40 I/O input/output
3GPP third generation partnership project K8 Kubernetes
4G fourth generation LCM lifecycle management
5G fifth generation LMF location management function
5GC 5G core network LTE long term evolution (4G)
ACF application component function 45 MAC medium access control
AMF access and mobility management func- MEAO MEC application orchestrator
tion MEC multi-access edge computing
API application programming interface MEC-CO MEC customer orchestrator
app application MEC-CVEO MEC customer VEO
APS application slice 50 MECO MEC customer orchestrator
APSMF application slice management function MEC-OO MEC owner orchestrator
APSS application slice subnet MEC-OVEO MEC owner VEO
APSSMF application slice subnet management MECPM MEC platform manager, or MEC custom-
function er platform manager
AS application service 55 MENSM MEC network slice manager
ASC application service consumer MEO MEC orchestrator
ASIC application-specific integrated circuit MEOO MEC owner orchestrator
ASMF application service management func- MEP MEC platform

22
 43 EP 4 178 169 A1 44

MEPM MEC platform manager requesting (2104), using the management web
MEPM-V MEC platform manager - NFV portal (524), the application service with a given
MME mobility management entity service level agreement from a catalogue of of-
MOP MEC operator fered application services; and
NCE network control element 5 communicating (2106), using the management
NF network function web portal (524), with an application service
NFV network function virtualization management function (504);
NFVI NFV infrastructure wherein the application service management
ng or NG new generation function (504) is configured (2108) to translate
ng-eNB new generation eNB 10 the service level agreement of the requested ap-
NG-RAN new generation radio access network plication service to a specification of an applica-
NR new radio (5G) tion slice, and to trigger a creation of an appli-
NSMF network slice management function cation slice instance with contacting an applica-
NSS network slice subnet tion slice management function (506).
NSSMF network slice subnet management func- 15
tion 2. The method (2100) of claim 1, further comprising:
NST network slice template manage at least one network service, the network
N/W or NW network service being related to the application service.
OS operating system
PDA personal digital assistant 20 3. The method (2100) of any of claims 1 to 2, wherein:
PDCP packet data convergence protocol
PHY physical layer the application slice instance is split with the ap-
PNF physical network function plication slice management function (506) into
QoS quality of service multiple subnets, one subnet of the multiple sub-
RAN radio access network 25 nets corresponding to a domain over which the
RLC radio link control application slice spans; and
RRC radio resource control (protocol) an application slice subnet management func-
RRH remote radio head tion (512, 522) applies an application slice sub-
RU radio unit net lifecycle management command with a do-
Rx receiver or reception 30 main relevant for a user equipment slice subnet,
SGW serving gateway and with a domain relevant for a multi-access
SLA service level agreement edge computing subnet.
SON self-organizing/optimizing network
TN transport network 4. The method (2100) of any of claims 1 to 3, further
TRP transmission and/or reception point 35 comprising:
TS technical specification partitioning the service level agreement in into a first
Tx transmitter or transmission service level agreement and a second service level
UE user equipment (e.g., a wireless, typical- agreement, the first service level agreement being
ly mobile device) independent of the second service level agreement.
UPF user plane function 40
VEO virtual environment orchestrator - e.g. 5. A non-transitory program storage device (1704)
OpenStack Heat, Kubernetes, etc. readable by a machine, tangibly embodying a pro-
VI virtualization infrastructure gram of instructions executable with the machine for
VNF virtualized/virtual network function performing operations, the operations comprising:
VNFM VNF LCM manager 45
X2 network interface between RAN nodes managing (2102) an application service using a
and between RAN and the core network management web portal (524);
Xn or XN network interface between NG-RAN requesting (2104), using the management web
nodes portal (802), the application service with a given
50 service level agreement from a catalogue of of-
fered application services; and
Claims communicating (2106), using the management
web portal (524), with an application service
1. A computer implemented method (2100) compris- management function (504);
ing: 55 wherein the application service management
function (504) is configured (2108) to translate
managing (2102) an application service using a the service level agreement of the requested ap-
management web portal (524); plication service to a specification of an applica-

23
 45 EP 4 178 169 A1 46

tion slice, and to trigger a creation of an appli-

cation slice instance with contacting an applica-
tion slice management function (506).

6. An apparatus (1700) comprising: 5

means for managing (2102) an application serv-

ice using a management web portal (524);
means for requesting (2104), using the manage-
ment web portal (802), the application service 10
with a given service level agreement from a cat-
alogue of offered application services; and
means for communicating (2106), using the
management web portal (802), with an applica-
tion service management function (524); 15
wherein the application service management
function (504) is configured (2108) to translate
the service level agreement of the requested ap-
plication service to a specification of an applica-
tion slice, and to trigger a creation of an appli- 20
cation slice instance with contacting an applica-
tion slice management function (506).

7. The apparatus (1700) of claim 6, further comprising

means for managing at least one network service, 25
the network service being related to the application
service.

8. The apparatus (1700) of any one of claims 6 to 7,

wherein: 30

the application slice instance is split with the ap-

plication slice management function (506) into
multiple subnets, one subnet of the multiple sub-
nets corresponding to a domain over which the 35
application slice spans; and
an application slice subnet management func-
tion (512 or 522) applies an application slice sub-
net lifecycle management command with a do-
main relevant for a user equipment slice subnet, 40
and with a domain relevant for a multi-access
edge computing subnet.

9. The apparatus (1700) of any one of claims 6 to 8,

further comprising means for partitioning the service 45
level agreement in into a first service level agreement
and a second service level agreement, the first serv-
ice level agreement being independent of the second
service level agreement.
50

55

24
EP 4 178 169 A1

25
EP 4 178 169 A1

26
EP 4 178 169 A1

27
EP 4 178 169 A1

28
EP 4 178 169 A1

29
EP 4 178 169 A1

30
EP 4 178 169 A1

31
EP 4 178 169 A1

32
EP 4 178 169 A1

33
EP 4 178 169 A1

34
EP 4 178 169 A1

35
EP 4 178 169 A1

36
EP 4 178 169 A1

37
EP 4 178 169 A1

38
EP 4 178 169 A1

39
EP 4 178 169 A1

40
EP 4 178 169 A1

41
EP 4 178 169 A1

42
EP 4 178 169 A1

43
EP 4 178 169 A1

44
EP 4 178 169 A1

45
EP 4 178 169 A1

46
EP 4 178 169 A1

47
EP 4 178 169 A1

48
 EP 4 178 169 A1

10

15

20

25

30

35

40

45

50

55

49
 EP 4 178 169 A1

10

15

20

25

30

35

40

45

50

55

50
 EP 4 178 169 A1

10

15

20

25

30

35

40

45

50

55

51
 EP 4 178 169 A1

REFERENCES CITED IN THE DESCRIPTION

This list of references cited by the applicant is for the reader’s convenience only. It does not form part of the European
patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be
excluded and the EPO disclaims all liability in this regard.

Non-patent literature cited in the description

• 3GPP, Management and orchestration; concepts, • 3GPP, Telecommunication management; Study on

use cases and requirements (release 17). Technical management and orchestration of network slicing for
Specification 28.530 V17.1.0, March 2021 [0058] next generation network (Release 15). Technical Re-
[0059] port 28.801 V15.1.0, January 2018 [0058]

52