5G Core Basic Concepts
5G Core Basic Concepts
1. The AMF performs access control and mobility management of the MME.
2. The SMF performs session management of the MME and integrates SGW-C and
PGW-C functions.
3. The PCF is equivalent to the PCRF.
4. The UDM is equivalent to the HSS.
5. The AUSF is equivalent to the 3GPP AAA server.
6. The UPF is equivalent to the SGW-U and PGW-U.
7. The NEF is equivalent to the SCEF. However, the exposure capabilities of the NEF do
not include the IMS part.
8. The NSSF is a new NF used to select network slices.
9. The NRF is a new NF used for service registration and discovery under the SBA.
Acronym:
AS Application Server
SCS Services Capability Server
In the LBO architecture, the PCF on the VPLMN may interact with the AF to generate
a PCC rule for the service delivered through the VPLMN. According to the roaming
agreement with the HPLMN operator, the PCF on the VPLMN uses the locally
configured policy as the input of PCC rule generation. The PCF on the VPLMN cannot
access the user policy information of the HPLMN.
23501 6.2.1.7
Security Edge Protection Proxy (SEPP)
The SEPP is a non-transparent proxy that supports the following functions:
- Inter-PLMN control-plane message filtering and forwarding
- Topology hiding
Compared with the 4G MME, the AMF does not provide the session management
function, but incorporates the NAS transparent transmission function and supports
non-3GPP access.
The UDM+UDR functions as the 4G HSS, but the UDR can store structured data for
the PCF and the NEF at the same time.
Trusted AFs can directly access an NF without communicating with the NEF.
An Application Function may influence UPF (re)selection and traffic routing via PCF or NEF
1. NRF, NNSF, AMF, SMF, UDM, AUSF, PCF, UPF, SMSF, and NEF
2. N1, N2, N3, N4, N6, or N9 interface (based on the reference point model)
NRF, NNSF, AMF, SMF, UDM, AUSF, PCF, UPF, SMSF, and NEF
N1, N2, N3, N4, N6, and N9 are interfaces based on reference points.
The NRF provides the NF registration, NF discovery, and service launch notification functions. These
functions help achieve NF interconnections and on-demand NF and service configurations. In the above
figures, NF service consumers include the AMF, SMF, and PCF.
Note: As specified in 3GPP R15, services are always associated with NFs. This way, a service must be
selected when an NF is selected, and cannot be selected independently.
NRF deployment:
PLMN level: The NRF is configured with all information about a PLMN. NFs and NF services for the
PLMN can be selected.
Shared slice level: The NRF is configured with information about multiple shared slices. NFs and NF
services for these shared slices can be selected.
Slice-specific level: The NRF is configured with the information about a specific slice. NFs and NF
services for this slice can be selected.
The preceding deployment modes are optional. Carriers can determine the layer where the NRF is to be
deployed based on actual requirements.
NF Services provided by the NRF
The "change of status" of the NFStatus service operations can imply a request to be
notified of newly registered NF Instances in NRF, or to be notified of profile changes
of a specific NF Instance, or to be notified of the deregistration of an NF Instance.
Data set ID
Indicates the data set to be supported by the NF to be discovered. May be included if
the target NF type is "UDR".
data sets (i.e. "/subscription-data", "/policy-data", "/exposure-data" and "/application-
data")
Reference:
23501 f40
7.2 Network Function Services
Service examples provided by the AMF (see section 5.2.2 in 3GPP TS 23.502):
Namf_Communication: The NF may communicate with a UE and the AN by
using the AMF. The SMF can request the EBI allocation for interworking with
the EPS.
Namf_EventExposure: Other NFs can subscribe to mobility-related events or
statistics from the AMF.
Namf_MT: The NF can use the service to check whether the UE is reachable.
For details about the services provided by each NF, see section 5.2 in 3GPP TS
23.502.
After receiving a session establishment request from the UE, AMF-1 requests the
NRF to discover a desired SMF. The NRF looks up the locally maintained NF
information, selects SMF-1, and sends SMF-1 information to AMF-1.
After obtaining SMF-1 information, AMF-1 invokes the corresponding service through
an SBI to establish a PDU session.
1. 5GC NFs interact based on the service-oriented architecture, and EPC NEs
interwork based on the reference point architecture. 5GC is more flexible. Each NF
has its own interface standard. Other NFS can interact with the NF as long as they
comply with this standard.
2. 5GC NFs are discovered by an NRF. Each NF registers its capabilities with the
NRF when going online. When other NFs send requests to the NRF, the NRF selects
proper NFs. Network adjustment is dynamic. EPC NE selection is based on static
configuration in the DNS.
The N1 interface is between the UE and AMF. It is used to transmit NAS messages between the UE
and core network. The NAS messages are irrelevant to the access layer.
The NAS can be classified into NAS-MM, NAS-SM, NAS-SMS, NAS-UE Policy, and NAS-LCS. NAS-
SM, NAS-SMS, NAS-UE Policy, and NAS-LCS are carried on NAS-MM.
When a UE accesses the AMF through 3GPP or non-3GPP access modes, each access mode has an
N1 NAS signaling connection.
NAS-MM supports the following functions:
Implements the AMF-terminated NAS procedure, including RM/CM status management, secure
NAS signaling connection, and access control for UEs and the AMF.
Transmits different types of NAS messages, such as NAS-SM and SMS, with the NAS-SM
messages at the same time.
Uses the same NAS protocol for 3GPP and non-3GPP access.
The security context between the UE and the AMF provides security functions for all NAS messages.
The AN includes the RAN for 3GPP access and the N3IWF for untrusted non-3GPP access. They use
the same NGAP protocol. The AMF is the only termination point of the N2 interface.
As the AMF is decoupled from other control plane functions, the NGAP allows the AMF to function as
the forwarding point of the 5G-AN and other control plane functions.
The N2 interface defines two different procedures:
Procedures related to the N2 interface management and are irrelevant to UEs, such as N2
interface setup or reset.
UE-related procedures, such as NAS message transmission, UE context management, PDU
session management, and handover management.
The destination well-known port of the PFCP Request message is 8805, and the
source port is allocated by the sender (specified in 3GPP TS 29.244).
When a UE is in the CM-IDLE state, the UPF can forward downlink data to the SMF.
The SMF triggers different paging policies according to the DSCP value. (Optional
procedure) (3GPP TS 23.502)
PDU layer: Indicates the PDU transmitted in the PDU session between a UE and a DN. When the
PDU session type is IPv4, IPv6, or IPv4v6, the PDU layer functions as IPv4 and IPv6 packets. When
the PDU session type is Ethernet, the PDU layer functions as an Ethernet frame.
GTP-U: This protocol is used to encapsulate multiple paths of different PDUs of user data on the N3
interface. A GTP-U tunnel corresponds to a PDU session, and the layer also carries an identifier
associated with a QoS flow.
GTP-U: This protocol is used to encapsulate the user data packets of N3 and N9 interfaces. A GTP-U
tunnel corresponds to a PDU session, and the layer carries the identifier associated with the QoS flow.
References
23003 2.2B Subscription Concealed Identifier (SUCI)
33501 6.12.2
As examples, assuming the IMSI 234150999999999, where MCC=234, MNC=15 and
MSISN=0999999999, the Routing Indicator 678, and a Home Network Public Key
Identifier of 27:
The SUCI for the null-scheme is composed of 0, 234, 15, 678, 0, 0 and 0999999999.
SIDF: UDM internal module. It can decrypt an SUCI to obtain the SUPI.
Reference: TS 33.501
The 5G HE AV is generated by the UDM and is the original authentication vector. The
5G AV is generated by the AUSF replacing the XRES in the 5G HE AV with the
HXRES.
HXRES = Hash (RAND||XRES), where HXRES in the 5G AV can be regarded as the
encrypted original XRES to improve the authentication security in visited areas.
Re-authentication is not required during handovers, and the anchor key is shared.
During a handover between a non-3GPP and a 3GPP network, the AMF is a unified
authentication anchor and does not need to obtain the authentication key again.
The UE is not registered with the network. Although the AMF may store some
UE contexts for next authentication, the contexts hold no valid location or
routing information for the UE.
The AMF can reject the registration request of the UE and maintain the current
state.
When the UE and the AMF are in the RM-Registered state, the UE can perform:
A deregistration procedure
A UE has a NAS signaling connection with the AMF, and has N2 and N3
connections.
The UE reachability (location) information in the AMF is the cell or base station
information.
Uplink and downlink data services and signaling procedures can be directly
performed.
When the UE/AMF is in the CM-IDLE state:
There is no NAS signaling connection between the UE and AMF, and the AN,
N2, and N3 connections do not exist.
When the AMF has data to be sent to the UE, the AMF sends a paging request
to the UE in the RA to execute the Service Request procedure triggered by the
network side.
When a UE changes from the CM-IDLE state to the CM-CONNECTED state, it needs to initiate a
service request procedure or a registration procedure (TAU), which consumes a lot of signaling. The
RRC Inactive state is introduced to quickly restore connections and reduce power consumption. In the
RRC Inactive state, the UE context is stored in the RAN. The N2 interface between the RAN and CN is
not released, but the RRC connection between the UE and gNodeB is released. In this case, the UE
enters the RRC Inactive state.
In the RRC Inactive state, the UE is in the CM-CONNECTED state from the perspective of CN.
The RAN determines whether a UE enters the RRC Inactive state. The CN provides the following
information for the RAN:
UE specific DRX values
RA provided to the UE
Periodic registration update timer
Whether the UE is in MICO mode
Information from the UE permanent identifier
When the UE is in the CM-CONNECTED with RRC Inactive state, the UE may resume the RRC
connection due to:
Uplink data pending
UE-initiated NAS signaling procedure
A response to RAN paging
Notification of the network that it has left the RAN-based notification area (RNA)
RAN area update procedure initiated when the RAN paging timer expires
The NG RAN triggers paging when:
Downlink NAS messages are to be transmitted.
Downlink data is to be transmitted.
If the UE resumes the connection in a different NG-RAN node within the same PLMN, the UE AS
context is retrieved from the old NG-RAN node, and a procedure is triggered towards the CN.
If the UE in the RRC Inactive state fails to restore the RRC connection, it enters the CM-IDLE state.
If the UE in the RRC Inactive state performs cell selection to the GERAN/UTRAN/E-UTRAN, it follows
idle mode procedures of the selected RAT. For details, see TS 23.501 5.17 (by 2018.5.26).
1. What are the similarities and differences between the 5GC UE status and the EPC
UE status?
Similarities: The MM-Registered and MM-Deregistered, as well as CM-
CONNECTED and CM-IDLE states are retained.
Differences: The CM-CONNECTED with RRC Inactive state has been added
for 5G to reduce the number of signaling interactions before a UE performs
data services.
2. In the CM-CONNECTED with RRC Inactive state, the downlink data can reach the
gNodeB, but the gNodeB cannot forward the data to the UE. Therefore, the gNodeB
must have the paging function. In addition, the P-TAU timer and the reachability timer
are not started. Therefore, the wireless side is required to provide the two functions.
Note: 3GPP TS 23.501 did not support IPv4v6 and stipulated that an IPv4 PDU
session and an IPv6 PDU session can be established at the same time if the
support for a dual-stack address is required. However, the latest version (3GPP TS
23.501-f10) supports IPv4v6 and stipulates the usage scenario: An IPv4v6 address
is applied for when the MT and TE are separated and the MT does not know the IP
address type supported by the TE. The SMF allocates IPv4, IPv6, and IPv4v6
addresses based on the DNN configuration and carrier policies. (By 2018-05-27)
The 5GC supports a PDU connectivity service (a service that provides exchange of
PDUs between a UE and a data network identified by a DNN). The PDU connectivity
service is supported using PDU sessions that are established upon requests from the
UE.
The SSC mode associated with a PDU session does not change during the lifetime
of a PDU session.
5G QoS is a framework based on QoS flows. The QoS flow is the finest granularity
for 5G QoS control.
The RAN can use multiple QoS flows to share one DRB according to the policy. For
example, GBR QoS flows use one DRB, and non-GBR QoS flows use one DRB.
Note: User-plane service policies (such as MFBR enforcement) are not regarded as
QoS differentiation. These policies are completed by UPFs on an SDF level
granularity.
5G QoS flow: The finest granularity for QoS forwarding treatment in the 5G system.
All packets mapped to the same 5G QoS flow receive the same forwarding
treatment (for example, scheduling policy, queue management policy, rate shaping
policy, and RLC configuration). Providing different QoS forwarding treatments
requires different 5G QoS flows.
If a QoS rule does not contain a packet filter set or a packet filter set contains all UL
packets, RQA cannot be enabled.
The UL and DL session-AMBRs are enforced by the UPF, and the DL session-AMBR
needs to be enforced separately in every UPF that terminates the N6 interface.
UL and DL UE-AMBRs are enforced by (R)AN.
If the UE receives a session-AMBR, it needs to enforce the UL session-AMBR.
NOTE 2: It is required that default MDBV is supported by a PLMN supporting the related 5QIs.
NOTE 3: This MDBV value is set to 1354 bytes to avoid IP fragmentation for the IPv6 based, IPSec protected GTP
tunnel to the 5G-AN node (the value is calculated as in Annex C of TS 23.060 [56] and further reduced by 4 bytes to
allow for the usage of a GTP-U extension header).
NOTE 4: A delay of 1 ms for the delay between a UPF terminating N6 and a 5G-AN should be subtracted from a given
PDB to derive the packet delay budget that applies to the radio interface.
NOTE 5: A delay of 2 ms for the delay between a UPF terminating N6 and a 5G-AN should be subtracted from a given
PDB to derive the packet delay budget that applies to the radio interface.
NOTE 6: A delay of 5 ms for the delay between a UPF terminating N6 and a 5G-AN should be subtracted from a given
PDB to derive the packet delay budget that applies to the radio interface.
23501 5.3.7
Priority Level
The Priority Level associated with 5G QoS characteristics indicates a priority in scheduling resources among
QoS Flows. The lowest Priority Level value corresponds to the highest priority.
Packet Delay Budget
The Packet Delay Budget (PDB) defines an upper bound for the time that a packet may be delayed between
the UE and the UPF that terminates the N6 interface.
Packet Error Rate
The Packet Error Rate (PER) defines an upper bound for the rate of PDUs (e.g. IP packets) that have been
processed by the sender of a link layer protocol (e.g. RLC in RAN of a 3GPP access) but that are not
successfully delivered by the corresponding receiver to the upper layer (e.g. PDCP in RAN of a 3GPP
access).
Default Averaging Window
Each GBR QoS flow is associated with an averaging window. The window indicates that the GFBR/MFBR
If RQ control is used, the RQA can be signaled to the RAN through the N2
interface, as described in clause 5.7.5.3. RQA indicates that certain traffic carried
on a QoS flow is subject to reflective QoS.
RQA is an optional parameter. It indicates that some (not all) traffic carried over
the QoS flow is affected by the RQA. The (R) AN allows RQI transmission for AN
resources corresponding to a QoS flow only if an RQA is signaled in the QoS
flow. The RQA may be sent to the NG-RAN via the N2 reference point at UE
context establishment in NG-RAN and at QoS Flow establishment or
modification.
Application scenario: The Reflective QoS mechanism can be used for a new SDF to be added to a QoS flow. The
downlink QoS rule is generated by the SMF and sent to the UPF, and the UPF then sends the service flows that
require UE derived QoS rule creation to the UE by labeling the service flows (with the RQI, FQI) on the user plane.
The uplink QoS rule is generated by the UE based on the downlink service flows to reduce the signaling for QoS
rule updates.
Control mechanism:
Control plane:
When the SMF determines to activate the reflective QoS mechanism, the SMF sends the SDF QoS control
information containing the RQI to the UPF, and sends a QoS profile containing the RQA to the AN.
User plane:
The UPF receives the data packets corresponding to the SDF, and contains the RQI in the tunnel of the
data packets.
The AN sets the air interface data packet header to contain the RQI according to the RQI in the data
packets.
When a UE receives data packets containing the RQI, the UE checks whether there are uplink QoS rules
corresponding to the data packets on the local end. If there are not, the UE generates a UE derived QoS
rule and starts a timer. If there are, the UE restarts a timer. When the timer expires, the UE derived QoS
rule is deleted. If there is an uplink QoS rule corresponding to data packets on the local end but the QFI
associated with the downlink packets is different from the QFI in the uplink QoS rule stored on the local
end, the UE updates the QFI corresponding to the QoS rule.
Exit mechanism
Control plane:
When the 5GC determines to no longer use reflective QoS for a specific SDF, the SMF
removes the RQI in the corresponding SDF information provided to the UPF through the
N4 interface.
User plane:
When the UPF receives this instruction for this SDF, the UPF no longer sets the RQI in
the header on the N3 reference point.
The UPF continues to accept the UL traffic of the SDF for the originally authorized QoS
flow for a carrier configurable time.
The SMF can control the data path of a PDU session so that the PDU session can correspond to
multiple N6 interfaces at the same time. Different UPFs of the same PDU session provide access to the
same DN. In addition, the UPF allocated during the PDU session establishment is associated with the
SSC mode of the PDU session, and the additional UPF (for example, the additional UPF used to be
selectively routed to a DN) allocated in the same PDU session is independent of the SSC mode of the
PDU session.
The selective DN service routing supports the forwarding of some selected services to the N6 interface
of a DN closer to the UE.
In the case of PDU sessions with the type of IPv4, IPv6, IPv4v6, or Ethernet, the SMF may decide to
insert a UL CL in the data path of a PDU session. The UPF diverts (local) traffic based on the UL CL
delivered by the SMF.
The insertion and removal of a UL CL are determined by the SMF. A UL CL is delivered over the N4
interface and executed by the UPF. The SMF may decide to insert a UPF supporting the UL CL function
in the data path of a PDU session during or after the PDU session establishment, or to remove a UPF
supporting the UL CL function from the data path of a PDU session after the PDU session
establishment. The SMF may insert more than one UPF supporting the UL CL function in the data path
of a PDU session.
The UPF supporting the UL CL function is used for uplink traffic steering and downlink traffic merging,
and can be used for traffic measurement, charging, and session-AMBR enforcement.
The UE is unaware of and does not participate in UL CL and traffic steering.
When the UPF supporting the UL CL function is inserted in the data path of a PDU session, there are
multiple PDU session anchors for the PDU session. These PDU session anchors (UPFs) provide
different access to the same DN. However, a PDU session of an IP address type has only one IP
anchor.
When there are multiple UPFs in the data path of a PDU session, only one UPF can provide the N3
interface.
One UPF can support both the UL CL and PDU session anchor functions.
A PDU session can be associated with multiple IPv6 prefixes. A UPF supporting the branching point
function can steer traffic to different anchors UPFs according to different UE IPv6 prefixes. The UE
selects different IPv6 prefixes for different services based on routing information and priorities. For
details about how the UE obtains routing information and priorities, see RFC 4191.
A UPF supporting the branching point function is used for uplink traffic steering and downlink traffic
merging, and can be used for traffic measurement, charging, and session-AMBR enforcement. The
insertion and removal of a UPF supporting the branching point function are determined by the SMF and
delivered through the N4 interface. The SMF may decide to insert a UPF supporting the branching point
function in the data path of a PDU session during or after the PDU session establishment, or to remove
a UPF supporting the branching point function from the data path of a PDU session after the PDU
session establishment.
IPv6 multi-homing takes effect only for IPv6. When the UE requests to establish an IPv6 or IPv4v6 PDU
session, the UE also indicates whether the UE supports the IPv6 multi-homing function.
The multi-homing PDU session can be used to support make-before-break service continuity to support
SSC mode 3. The multi-homing PDU session can also be used to support cases where a UE needs to
access both a local service (such as a local server) and a central service (such as the Internet).
A UPF can support both the branching point and the PDU session anchor functions.
The UE has two IPV6 addresses in this case.
An LADN service area is a set of tracking areas. The LADN service is provided by the visited PLMN.
The LADN service applies only to 3GPP networks and does not apply to the home routed
scenario.
The LADN service is a part of the DNN subscription data.
The UE needs to know whether the DNN is an LADN DNN.
The LADN information is delivered by the AMF to the UE during the registration procedure or UE
Configuration Update procedure. The AMF delivers only the TA (that is, the intersection of the LADN
service area and the current registration area) that belongs to the current registration area of the UE.
Based on the LADN information in the UE, the UE performs the following actions:
(a) When the UE is out of an LADN service area, the UE:
Must not request to activate the user plane connection of a PDU session for this LADN
DNN.
Must not request to establish or modify a PDU session for this LADN DNN.
Does not need to release any existing PDU session for this LADN DNN unless the UE
receives explicit PDU session release request from a network.
(b) When the UE is in an LADN service area, the UE:
May request a PDU session establishment or modification for this LADN DNN.
May request to activate the user plane connection of the existing PDU session for this
LADN DNN.
The following parameter(s) and information may be considered by the SMF for UPF selection and re-
selection:
- UPF's dynamic load.
- UPF's relative static capacity among UPFs supporting the same DNN.
- UPF location available at the SMF.
- UE location information.
- Capability of the UPF and the functionality required for the particular UE session: An appropriate UPF can
be selected by matching the functionality and features required for an UE.
- Data Network Name (DNN).
- PDU Session Type (i.e. IPv4, IPv6, IPv4v6, Ethernet Type or Unstructured Type) and if applicable, the
static IP address/prefix.
- SSC mode selected for the PDU Session.
- UE subscription profile in UDM.
- DNAI as included in the PCC Rules and described in clause 5.6.7.
- Local operator policies.
- S-NSSAI.
- Access technology being used by the UE.
- Information related to user plane topology and user plane terminations, that may be deduced from:
- AN-provided identities (e.g. CellID, TAI), available UPF(s) and DNAI(s);
- Information regarding the user plane interfaces of UPF(s).This information may be acquired by the SMF
using N4;
- Information regarding the N3 User Plane termination(s) of the AN serving the UE. This may be deduced
from AN-provided identities (e.g. CellID, TAI);
- Information regarding the N9 User Plane termination(s) of UPF(s) if needed;
-Information regarding the User plane termination(s) corresponding to DNAI(s).
Network slice is an E2E network, including the RAN, transport network (TN), and core network (CN). It
requires a cross-domain slice management system.
Network slices require isolation of resources, security, and OAM. Different domains can use different
technologies to achieve this isolation. For example, the CN uses virtualization technology. FlexE and
VPN technologies can be used to provide better service isolation with a given amount of bandwidth. The
RAN can use differentiated air interface technologies for resource isolation.
Slices A, B, C are typical network slices. Each slice has two instances.
The NFs in slice A are completely isolated.
Some control plane NFs are shared between instances, and some are isolated. User plane
NFs are isolated on slice B.
The entire control plane is shared, and the entire user plane is isolated on slice C.
There are three standard SSTs: eMBB (1), URLLC (2), and MIoT (3).
Example:
0x01000001 identifies a slice of the eMBB type. There may be three slice instances for selection.
0x02000001 and 0x02000002 identify a slice of the URLLC type and share one slice instance.
1. If a UE attempts to register with a PLMN using a radio access technology (RAT), the UE first sends a
Registration Request message to the (R)AN. If the UE stores the configured NSSAI of the PLMN or the
allowed NSSAI for the RAT and the PLMN, the UE includes the requested NSSAI in the NAS
Registration Request message and the AN message. The requested NSSAI includes the S-NSSAI of the
UE-requested slice.
2. The (R)AN selects the initial AMF based on the globally unique AMF ID (GUAMI) or requested NSSAI.
If the UE does not provide the requested NSSAI or GUAMI in the AN message, the (R)AN sends the
Registration Request message from the UE to the initial AMF.
3. The initial AMF looks up the UDM for UE subscription data including the subscribed S-NSSAIs. The
initial AMF determines whether it can provide services for the UE, according to the received requested
NSSAI, subscribed S-NSSAI, and local configuration. If the initial AMF can serve the UE, it is the serving
AMF. It constructs the allowed NSSAI based on the subscribed S-NSSAI and the requested NSSAI and
sends the allowed NSSAI to the UE through a registration accept message. If the initial AMF cannot
provide services for the UE or cannot make a decision, the AMF needs to send a query request to the
NSSF for information about the target AMF.
4. Query parameters include the requested NSSAI, subscribed S-NSSAI, SUPI, and TAI of the UE.
5. The NSSF selects an AMF set or a candidate AMF list that can serve the UE, and an allowed NSSAI
suitable for the RAT. It may further select an NSI serving the UE, and an NRF used to select an NF in
the instance. These are selected according to the received information and local configuration. The
information is sent to the initial AMF after selection.
6. If the initial AMF is not in the AMF set or the AMF address is unavailable in local
information, the initial AMF obtains the candidate AMF list by checking with the NRF.
The NRF sends a group of available AMF lists, including the AMF pointer and
address. The initial AMF selects an AMF as the target AMF. If the initial AMF cannot
obtain the candidate AMF list by checking with the NRF, the initial AMF needs to send
the Registration Request message to the target AMF through the (R)AN. The
message includes the AMF set and the allowed NSSAI.
7a. If the initial AMF directly sends the NAS message to the target AMF based on
the local policy and the subscription data, the initial AMF sends the Registration
Request message and other information except for the AMF set obtained from the
NSSF to the target AMF.
7b. If the initial AMF determines to forward the NAS message to the target AMF
through the (R)AN based on the local policy and subscription information, the initial
AMF sends a Reroute NAS message to the (R)AN. The Reroute NAS message
includes the target AMF set information and the Registration Request message, and
the related information obtained from the NSSF.
8. After receiving the registration request message sent in step 7, the target AMF
continues with the registration procedure. It sends the UE a registration accept
message containing the allowed NSSAI and NSSP.
Session management:
The control planes for session management on different slices are isolated.
The AMF and SMF are decoupled. Different SMFs can be selected for different slices.
The user planes for session management on different slices are isolated.
The UPFs for different slices can be independent from each other.
A session cannot be migrated between different slices.
The SMF sends the S-NSSAI mapping to a session to a gNodeB, so that the
gNodeB can allocate resources based on the S-NSSAI.
A UE selects the HPLMN S-NSSAI mapped to the session according to the
configured NSSP.
In roaming scenarios, the UE maps the HPLMN S-NSSAI to the VPLMN S-
NSSAI based on the mapping.
1. The AMFs can be shared, but the SMF and UPF cannot.
2. A maximum of eight slices.
1. The AMF requests the NSSF to provide the NRF information based on the S-
NSSAI.
2. The NSSF sends the NRF information.
3. The AMF requests the NRF to provide the SMF information.
4. The NRF sends the SMF information.
Note: If the AMF has obtained the NRF information, steps 1 and 2 can be skipped. (For
example, the AMF has obtained the NRF information during the attach procedure.)
1. The AMF requests the NSSF to return the information about the NRF used for SMF
selection based on the S-NSSAI.
2. The NSSF returns the NRF information.
3. The AMF requests the NRF to return the SMF information.
4. The NRF returns the SMF information.
Note: If the AMF has identified the NRF, steps 1 and 2 can be skipped. (For example, the
NRF information has been obtained during the attach procedure.)
1. PDU sessions are similar to PDN connections. QoS flows are similar to bearers.
QoS rules are similar to TFTs.
2. The Notification Control and RQA have been added for the 5GC. The changes are
made to reduce the signaling resources required for creating dedicated bearers and
updating the TFTs.
8a-c. If the AMF is different from the AMF registered previously, the SUPI provided by the UE cannot indicate a valid context in the
AMF, or the UE registers with the AMF registered in non-3GPP access, the new AMF registers with the UDM and subscribes to
UDM deregistration notification for the AMF. The UDM stores the AMF identifier and the associated access type. The AMF
obtains the access and mobility subscription data from the UDM, and the SMF selects the subscription data. The new AMF
provides the access type of the serving UE for the UDM, and the access type is set to 3GPP access. The UDM stores the
associated access type and serving AMF in the UDR. After obtaining the mobility subscription context from the UDM, the new
AMF sets up an MM context for the UE and subscribes to related status from the UDM.
8d. When the UDM stores the associated access type and the serving AMF, as described in step 8a, the UDM initiates a
Nudm_UECM_DeregistrationNotification service operation to the old AMF corresponding to the 3GPP access. The old AMF
removes the MM context of the UE. If the service NF removal reason indicated by the UDM is the initial registration, the old
AMF notifies all the UE-associated SMFs that the UE deregisters from the old AMF, and the old AMF unsubscribes from the
subscription information in the UDM.
9. If the AMF decides to communicate with a PCF, the AMF selects the PCF. If the new AMF obtains the PCF ID from the old AMF
in step 2 and is successfully associated with the PCF identified by the PCF ID, the AMF selects the (V-) PCF through the PCF
ID. If the PCF identified by PCF ID is unavailable or the PCF ID cannot be obtained from the old AMF, the AMF selects the
PCF.
10. The PCF may request UE event subscription.
11. (Optional) If the registration request in step 1 contains the PDU session to be activated, the AMF requests the SMF related to
the PDU session to activate the user plane connection of the PDU session. If the PDU session status indicates that the PDU
session is released by the UE, the AMF instructs the SMF to release the network resources related to the PDU session.
12. The new AMF sends a registration accept message (5G-GUTI, RA, mobility restriction information, PDU session status, allowed
NSSAI, periodic registration update timer, LADN information, accepted MICO mode, IMS voice over PS session supported
indication, and emergency service support indicator) to notify the UE that the registration request is accepted.
13. (Optional) If a new 5G-GUTI is allocated, the UE sends a registration complete message to the AMF for confirmation.
1. A UE requests to establish a PDU session. The request message contains slice information, DNN, SSC
mode, and PDU type. If this procedure is used for inter-system handovers or 3GPP and non-3GPP
handovers, this message also contains the "Existing PDU Session" indication.
2. The AMF selects an SMF for PDU session establishment based on information, including the slice
information and DNN. If the request message contains the "Existing PDU Session" indication, the AMF
selects an SMF based on the mapping between the PDU session ID or DNN (saved on the HSS and UDM)
and the SMF.
3. The SMF obtains session-related subscription data from the UDM.
5. The SMF obtains session-related policy information from the PCF.
6-12. The SMF selects a UPF based on information, including the UE location, DNN, and S-NSSAI. An E2E
user-plane path is set up. During this procedure, the SMF or UPF allocates uplink tunnels and notifies the
RAN of such information, and the RAN allocates downlink tunnels and notifies the UPF of such information.
13. The SMF subscribes to UE mobility event notifications on the AMF. For example, a notification is sent if
a UE moves to or out of the specified service area.
1. A UE requests to establish a PDU session. The request message contains slice information, DNN, SSC
mode, PDU type, and UE encapsulation location on the RAN. If this procedure is used for inter-system
handovers or 3GPP and non-3GPP handovers, this message also contains the "Existing PDU Session"
indication.
2. The AMF selects an SMF for PDU session establishment based on information, including the slice
information and DNN. If the request message contains the "Existing PDU Session" indication, the AMF
selects an SMF based on the mapping between the PDU session ID or DNN (saved on the HSS and UDM)
and the SMF.
3. The SMF obtains session-related subscription data from the UDM, and obtains session-related policy
information from the PCF.
5. The SMF selects a UPF based on information, including the UE location, DNN, and S-NSSAI.
6-15. An E2E user-plane path is set up. During this procedure, the SMF or UPF allocates uplink tunnels
and notifies the RAN of such information, and the RAN allocates downlink tunnels and notifies the UPF of
such information.
16. The SMF subscribes to UE mobility event notifications on the AMF. For example, a notification is sent if
a UE moves to or out of the specified service area.
Answer to question 1: A main difference between the 5GC registration procedure and
the EPC attach procedure is that a session cannot be established in the 5G
registration procedure, and a PDN connection must be established in the EPC attach
procedure. In addition, slices are introduced in the 5GC registration procedure, and
therefore a network slice that the UE is allowed to access is determined in the 5GC
registration procedure, the gNodeB also selects an AMF based on the slice that the
UE wants to access.
Answer to question 2: The 5GC session establishment procedure is similar to the
EPC PDN connection establishment procedure. The difference is that the SMF and
UPF need to be selected based on slice information. In addition, the 5GC supports
different SSC modes. The 4G network does not have the SSC mode concept. Only
one tunnel is established for 5G PDU sessions, and one tunnel is established for each
bearer of the 4G PDN connection.