lte_session_management
lte_session_management
Disclaimer
The contents of this document are subject to revision without notice due to
continued progress in methodology, design and manufacturing. Ericsson shall
have no liability for any error or damage of any kind resulting from the use of this
document.
Contents
6 PDN Disconnection 45
6.1 UE- or MME-Initiated PDN Disconnection 45
6.1.1 Traffic Case 45
6.2 PGW-Initiated PDN Disconnection 47
6.2.1 Traffic Case 48
10 Modification of Bearers 65
10.1 PGW-Initiated Bearer Modification without Bearer QoS
Update 65
10.1.1 Traffic Case 65
10.2 PGW-Initiated Bearer Modification with Bearer QoS Update 66
10.2.1 Traffic Case 66
10.3 HSS-Initiated Subscribed QoS Modification 68
10.3.1 HSS-Initiated Subscribed UE-AMBR Modification 68
10.3.2 HSS-Initiated Subscribed APN-AMBR or Bearer Level QoS
Modification 69
10.4 QoS Modification in IRAT Mobility 71
10.4.1 Traffic Case 71
10.5 Establishing an S11-U Connection for Data over NAS 72
10.6 E-UTRAN Initiated E-RAB Modification Procedure 73
14 Compliance 79
The EPS session management function of the MME establishes and handles the
connections between the UE and one or several PDNs.
A PDN connection consists of one default bearer and, depending on the service
for which the connection is used, several dedicated bearers. Each default or
dedicated bearer is assigned a payload tunnel between the UE and the PGW,
through the eNodeB and the SGW. A UE can have one or several PDNs.
PDN Connection
Dedicated bearer
Default bearer
The default bearer remains established throughout the lifetime of the PDN
connection to provide the UE with IP connectivity to the PDN. The number of
dedicated bearers that are established can vary throughout the lifetime of the
connection. The information about the PDN connection is stored in the MME. This
information is used when connections or bearers are for instance, modified or
deactivated.
The default bearer is a non-GBR bearer. A dedicated bearer can be either a GBR
or a non-GBR bearer.
For more information about the EMM and ECM states, see LTE Mobility
Management.
The EPS Session management function handles the following main procedures:
— Modification of Bearers
— S1 Release
For more information about the Mobility Management procedures that involve
Session Management, see LTE Mobility Management.
For information about session management in the S4-SGSN, see GSM and
WCDMA Session Management.
Each GBR bearer is also associated with the following bearer-level QoS
parameters:
MBR The MBR parameter limits the bit rate provided by a GBR
bearer. For example, excess traffic can be discarded by a
rate shaping function. The MBR can be greater than or
equal to the GBR for a particular GBR bearer.
1.3 AMBR
The AMBR is described using the following parameters:
The MME now supports sending a timestamp and maximum waiting time to the
SGW in the Create Session Request message. The timestamp is used by the
SGW or other network functions to ensure that the most recent PDN session is
not incorrectly overwritten. Together with the maximum waiting time, the
timestamp is also used by the SGW or other network functions to stop further
processing of Create Session Request messages which have already timed
out at the MME. This can be achieved by optionally setting the new
SendOriginationTimeStamp parameter to on and the new MaxWaitTimeOffset
parameter to 0 or a positive integer smaller than the value of
T3ResponseCreateSessionMs and T3ResponseCreateSessionMsS8,
respectively. However, this optional function is sensitive to the accuracy of the
network time and to the time synchronization between network elements. If time
synchronization between network elements or the accuracy of the network time
cannot be guaranteed, Ericsson recommends disabling this function by setting
the SendOriginationTimeStamp parameter to off.
The Time Zones function sends the UE local time zone to the SGW, which enables
time-based charging in the SGW or PGW, or both. For more information about
the Time Zones function, refer to LTE Mobility Management.
For UEs not supporting EMM-REGISTERED without PDN Connection, the initial
PDN connection is established during the Attach procedure.
For UEs supporting EMM-REGISTERED without PDN Connection, the initial PDN
connection can be created through the Attach procedure or the Standalone PDN
Connectivity procedure.
For information about the Attach procedure, see LTE Mobility Management and
Inter-System Mobility Management.
One or several PDNs can be accessed by using the same PGW or different PGWs.
It is only possible to have one SGW per UE.
If the UE is in the ECM-IDLE state, the Service Request procedure or the Control
Plane Service Request procedure is performed before the PDN Connectivity
procedure can be initiated. The Service Request procedure is described in LTE
Mobility Management. The Control Plane Service Request procedure is described
in Massive IoT.
4.1 E-RAB Setup Request (Activate Default EPS Bearer Context Request)
/ Initial Context Setup Request (Activate Default EPS Bearer Context Request)
4.2 RRC Connection Reconfiguration (Activate Default EPS Bearer Context Request)
5. Notify Request
6. Notify Answer
If the resolved APN matches the configured APN blacklist for the UE, the
PDN Connectivity Request is rejected with the ESM cause code #32 Service
option not supported.
2. The MME selects an SGW, if the request is to establish the first PDN
connection for the UE in the EMM-REGISTERED state without PDN
Connection.
The MME uses the subscription data to verify that the APN that is requested
by the UE is allowed and determines which PGW to contact. For more
information about APN resolving and PGW selection, see PGW and SGW
Selection during PDN Connectivity Procedures on page 12. The MME
allocates the identity bearer and sends a Create Session Request
message to the PGW through the SGW. The request message includes the
International Mobile Subscriber Identity (IMSI) of the UE, an APN, and QoS
parameters, and the TEID of the MME for signaling traffic.
The MME can set the S11-U Tunnel flag and the Control Plane Only PDN
Connection Indication flag and include the S11-U MME F-TEID in the
Create Session Request message according to the MME local
configuration policies, the UE-preferred CIoT network behavior, and the UE-
supported CIoT network behavior. For more information, see Massive IoT.
3. The PGW creates an entry in the EPS bearer context table. After negotiating
with the PCRF, the PGW sends a Create Session Response message to
the MME through the SGW.
The MME can set the S11-U Tunnel flag and the Control Plane Only
PDN Connection Indication flag and include the S11-U MME F-TEID
in the Create Session Request message according to the MME
local configuration policies, the UE-preferred CIoT network behavior
and the UE-supported CIoT network behavior. For more information,
see Massive IoT.
5. If the UE has non-3GPP capability, and the PGW identity differs from the
one that the HSS indicated in the subscription data, the MME sends a Notify
Request message containing the PGW identity and the APN to the HSS. If
the PDN connection is for the emergency service, only the emergency PGW
identity is sent in the Notify Request message.
The MME can be configured to send fewer Notify Request messages to the
HSS. For more information, see Subscriber Data Management.
7. If the request described in establishes the first SGi PDN connection for the
UE, and the TAI list changes as a result based on the selected SGW, the MME
sends the GUTI Reallocation Command message with an updated TAI list
to the UE.
When the MME receives the APN_CONGESTION cause code and a Back-Off Time
IE from the PGW, the MME performs the maximum number of fallback attempts
specified by the value of the GwMaxAttemptNumber parameter minus 1 to the
PGWs in the gateway candidate list. If all the tried PGWs indicate APN
congestion and back-off time, the lowest back-off time value is sent to the UE.
During congestion, the MME stores the back-off time value included in the
Create Session Response message sent from the PGW for a specific APN.
The MME rejects the Attach or PDN Connectivity procedures without sending a
Create Session Request message, if all the PGWs indicate congestion and
have an active back-off timer for the specific APN.
The APN-Based Congestion Control mechanism does not affect UEs that are
MPS-classified in the subscription data or UEs that indicate emergency in the
Attach Request and PDN Connectivity Request messages.
During PDN connectivity procedures, the MME selects a PGW and an SGW.
PGW Selection SGW Selection
imsi_and_ga_based_
Off
Candidates not exist sgw_selection
parameter
On
Retrieving PGWs from IMSI-Based
Candidates exist
Gatewa Selection Profile
ImsiGaBasedSgwSelection
Off
Candidates not exist parameter on IMSINS level
On
Retrieving PGWs
Candidates exist
from SAPC over Smp
Retrieving SGWs Based on
Candidates exist
IMSI Number Series and GA
Candidates not exist
Candidates not exist
On On
Off Off
Retrieving PGWs Retrieving PGWs Retrieving SGWs Retrieving SGWs
Candidates exist Based on DNS from Static PGW Candidates exist Based on DNS from Static SGW
Query with APN Selection Table Query with TAI Selection Table
Retrieving PGWs Based on Retrieving SGWs Based on
Candidates not exist Candidates not exist Candidates exist Candidates not exist Candidates not exist Candidates exist
DNS Query with APN DNS Query with TAI
gtpc_network_load_control gtpc_network_load_control
and Off
and Off
GwSelectionWithGtpcLocadInfo GwSelectionWithGtpcLocadInfo
parameters parameters
On On
Figure 3 PGW and SGW Selection during Attach and PDN Connectivity
Procedures
The PGW selection function of the MME allocates a PGW to provide PDN
connectivity for 3GPP access. The UE requests the APN and the subscriber
information provided by the HSS. An APN is a logical name referring to PDN
connections and the service to which the subscriber connects. The MME can also
use configured selection criteria when selecting the PGW.
2. If it exists, the MME uses a static PGW in subscription data from the HSS.
5. If static PGW selection is enabled on the MME before a DNS query, the MME
selects a locally configured PGW.
For IMS Emergency PDN, the emergency PGW selection mode can be configured
as local preferred or DNS preferred:
— If the emergency PGW selection mode is local preferred, the emergency PGW
lists from the MME emergency configuration data are used. In this mode, the
emergency PGW cannot be configured in the local static PGW table, but only
be configured in the emergency profile. When multiple PGWs have the same
priority in the emergency profile, the MME selects the PGW randomly.
— If the emergency PGW selection mode is DNS preferred, the PGW selection
based on the APN in the MME emergency configuration data is performed to
DNS.
If an IMS emergency PDN handed over from non-3GPP access is used for the
non-roaming authenticated UE, the emergency PGW from the MME emergency
configuration data, the emergency PGW from the DNS resource record, or the
emergency PGW from the HSS is used. For more information, see Configuring
MMTel Service.
The PGW is not changed during mobility procedures. However, DNS queries can
be performed to collect additional information related to existing PDN
connections.
3. If static SGW selection is enabled on the MME before a DNS query, the MME
selects a locally configured SGW.
4. Otherwise, the MME uses DNS queries to select the SGW based on TAI.
For the SGW Reselection procedure during mobility procedures, see section
Specific Gateway Selection during Mobility Procedures on page 38.
After the PGW and the SGW candidates are selected, the MME continues the
gateway selection procedure by sorting the gateway pairs and retrieving their IP
addresses.
PGW and SGW candidate selections are independent of each other until the
sorting of gateway pairs. Therefore, it is possible to use a different GW selection
method for each candidate. For example, the PGW selection can be based on
static configuration and the SGW selection can be based on the DNS query, or the
other way around. However, Ericsson recommends maintaining uniformity and
using the same selection method for the PGW and the SGW.
Note: The DECOR feature provides the possibility to select dedicated SGWs
and PGWs serving specific types of UEs. For more information, see
Dedicated Core Network.
The purpose of APN resolving is to determine which APN name to use for a PDN
connection.
— If the UE does not request an APN, the default APN in the subscription data
is used.
— If a DNS query fails or the UE requests an APN that does not match any APN
in the subscription data, the PDN connection is rejected.
For the Attach combined with PDN Connectivity Request procedure, if the
resolved APN matches the configured APN blacklist, the MME rejects the PDN
Connectivity Request and the Attach Procedure based on whether the UE
supports EMM-REGISTERED without PDN Connection. For more information, see
APN Resolve and Redirect for LTE Access.
After the formulation of the APN name, the MME uses the APN to select a PGW
during the configured IMSI-based or static PGW selection or to select a PGW
during DNS query with APN.
The SGSN-MME selects a PGW based on the IMSI or IMSI number series when
the IMSI-Based Gateway Selection function is enabled and the following
configurations are performed:
— The PGW list is configured with APN and PGW combinations. Each APN and
PGW combination consists of an APN, a primary PGW, and optionally a
secondary PGW for the APN.
For a specific UE with IMSI, the MME uses the APN for PDN connection setup
and as the input to find a PGW from the configured PGW list. If the APN does not
match any configured specific APN and a wildcard APN is configured, then the
wildcard APN is used as the input to find a PGW from the configured PGW list.
IMSI-based PGW selection is applicable for roaming users that set up a PDN
connection with local breakout and for home users.
For more information about how to configure IMSI-based PGW selection, see
Configuring IMSI-Based Gateway Selection.
Note: IMSI-based PGW selection is applicable for the S4-SGSN and the MME.
For more information about PGW selection with the Mobility-Based Policy
feature, see Mobility-Based Policy.
The MME selects a PGW from the configured static PGW selection tables in the
MME when the APN table, the PGW selection table, and the PGW table are
configured.
The MME uses the APN as the input to find candidate PGWs from the static PGW
selection tables. When the Advanced 4G/5G Interworking feature, the Dedicated
Core Network feature, and the Gateway Selection for NR Usage feature are
enabled, the MME uses the SMF, the UUT, and the NR Capability as extra filters to
select a PGW from the candidate PGWs. For more information about PGW
selection using the SMF, the UUT, and the NR Capability, see Advanced 4G/5G
Interworking, Dedicated Core Network and 5G EPC. The APN-OI replacement
functionality can also be used in the static PGW selection. For more information
about APN-OI replacement, see APN Resolve and Redirect for LTE Access.
Static gateway selection can be used to select the PGW for a roaming user only if
the roaming user sets up a PDN connection with local breakout.
The MME continues the gateway selection by using an internal FQDN based on
the configured PGW hostnames to sort the PGW candidates, see Sorting PGW
Candidates on page 20.
The DNS query uses the resolved APN to find the PGW candidates, see DNS S-
NAPTR Procedure for the PGW on page 17. For more information about APN
resolving, see APN Resolving on page 14.
The MME continues by sorting the PGW candidates, see Sorting PGW Candidates
on page 20.
When one, or both, of the two features Dedicated Core Network and Gateway
Selection for NR Usage are activated, multiple DNS S-NAPTR procedures are
performed with additional filtering.
When the Gateway Selection for NR Usage feature is activated and the UE is
allowed to use NR as a secondary RAT, the following two S-NAPTR procedures
are performed:
The PGW candidates are sorted as described in section 4.12.1. For more
information about the Gateway Selection for NR Usage feature, see 5G EPC.
When the Dedicated Core Network feature is activated, the following two S-
NAPTR procedures are performed:
The PGW candidates are sorted as described in section 4.12.2. For more
information about the Dedicated Core Network feature, see Dedicated Core
Network.
Note: When both features are activated and the UE is allowed to use NR as a
secondary RAT, an additional S-NAPTR procedure is performed. This
procedure retrieves the PGW candidates supporting NR which are
dedicated to the specific UUT. For example, the filtering candidates
include both +nc-nr and +ue-<uut> appended to the protocol. These
candidates have the highest priority when the PGW-SGW pairs are
sorted.
The services used depend on the applicable interface and protocol and are listed
as follows:
— x-s5-gtp
— x-s5-pmip
— x-s8-gtp
— x-s8-pmip
— x-gp
— To prepare 5GC combined PGW-C and SMF nodes for 5GC network
deployment, the MME prioritizes to select a combined PGW-C and SMF node
for a UE that has 5GS subscription, regardless of the N1 mode capability.
Due to the flexible configuration in the MME per IMSI number series, the MME
provides four PGW selection strategies for LTE UEs, regardless of UE 5GS
subscription support or N1 capability. The four PGW selection strategies are as
followings:
— PGW only: Only select a standalone PGW
— SMF ignored: Ignore the nc-smf service parameter in the gateway selection
In addition, the MME provides another strategy for the SGW relocation. In the
mobility with SGW relocation procedure, the MME only selects an SGW for the UE
and skips the SMF rule.
For the action configured to prioritize to select a combined PGW-C and SMF, the
MME prioritizes to select combined PGW-C and SMF nodes when establishing
PDN connections. The following two S-NAPTR procedures are performed:
For more information about gateway selection order based on SMF rules, see
Gateway Selection Based on Service Parameters.
For more information about the S-NAPTR procedure, see DNS Description.
<"topon" | "topoff">.<single-label>.<canonical-node-name>
Note: If both the topon and topoff labels are missing in an FQDN, it is
assumed that a topological structure is not used (topoff).
The sorting algorithm in the MME sorts the gateway candidates according the
following:
— If a configured static PGW exists, the MME sorts the candidate gateways
according to the gateway capacities.
— If the SAPC provides a PGW list, the MME selects the gateway candidates
randomly.
The SRV records from the DNS are sorted according to priority, lowest priority
first. Furthermore, each group of SRVs with equal priority is sorted according to
weight. For example, if record A has weight 20 and record B has weight 10,
record A is prioritized before B in 67% of the cases.
NAPTR records from the DNS are sorted according to order: lowest order first.
Also, each group of NAPTR records with the same order is sorted according to
weight, with the same algorithm as for SRVs. The difference is that the weight is
calculated out of the default base preference value of 65,535 in the algorithm.
For example, if record A has a preference value of 65,515 and record B has a
preference value of 65,525, the value 65,535 - 65,515 = 20 is used as relative
weight for A. The value 65,535 - 65,525 = 10 is used as relative weight for B. As a
result, A is prioritized before B in 67% of the cases.
Using the relative weight to sort the NAPTR and SRV records results in load
sharing between the supported PGWs.
3.7.1 Optional Function for Sorting the PGW with the GTP-C Load Information
The Gateway Selection with GTP-C Load Information function enables the MME
to select the PGW based on the GTP-C Load Control Information (LCI). This
function helps to achieve a balanced load network by using the dynamic load
information provided in the Load Control Information IE over the S11
interface.
Note: This feature is supported for the home subscribers and roaming
subscribers that are allowed by subscription data and local
configuration to use local breakout for all APNs. For the roaming users
with home routed traffic, the GTP-C Network Load Control function is
not applied.
This feature is activated based on local configuration per PLMN. For information
about the configuration of gateway selection based on the LCI, see Configuring
Session and Mobility Management.
When the Gateway Selection with GTP-C Load Information function is activated,
the MME takes into account the node level LCI received from the PDN GW in
addition to the weight factors received in the DNS S-NAPTR procedure to select
an appropriate PDN GW.
To ensure an efficient load control in the network, the MME updates the load
control information if the new received load metric has an increase or decrease
that is more than or equal to the configured value of
LoadControlInfoUpdateThreshold.
During the gateway selection procedure, the MME computes the dynamic weight
factor for the candidate gateways within the same order and priority in the SRV
records and uses the dynamic weight factor instead of the weight factor received
from the DNS server to sort the NAPTR and SRV records results in load sharing
among the supported PGWs. Considering the current available load and DNS
weight-factor of the target node as follows.
For NAPTR procedure, calculate the dynamic weight factor by the following
formula:
For SRV procedure, calculate the dynamic weight factor by the following formula:
Note: If the load metric is not available, the MME considers the load metric as
0 for a candidate target node.
Example:
This assumes that the procedures specified in this document have 3 candidate
PGWs with the same relative order (e.g. with the same topological order) and the
following weights are received from the DNS and load metric reported via GTP-C
signalling:
Based on the formula above, the selecting node calculates the dynamic weight
factor for each candidate PGW:
The MME supports the LCI IE in the below messages sent from the PGW:
— The primary SGW and optionally the secondary SGW are configured in the
profile.
IMSI-based SGW selection can be used for selecting the SGW for emergency UEs
if any IMSI exists.
Note: IMSI-based SGW selection is applicable for the S4-SGSN and the MME.
The SGSN-MME continues the gateway selection procedure by sorting the PGW-
SGW pairs. For more details, see Sorting PGW-SGW Pairs on page 28.
For information about SGW selection during mobility procedures when IMSI-
based SGW selection is configured, see IMSI-Based Gateway Selection on page
39.
When additional PDN connections are established using the PDN connectivity
procedure, the SGW is not changed.
The MME selects an SGW from the SGW table when the following configurations
are performed:
— In the SGW table, the SGW configuration entry with IMSINS reference ID list
is configured.
The MME uses the Tracking Area Identifier (TAI) and IMSI number series ID as
the input to find candidate SGWs from the static SGW selection table. When the
Dedicated Core Network feature and Gateway Selection for NR Usage feature,
are enabled, the MME uses UUT and NR Capability as the extra filter to select
SGW from the candidate SGWs. For more information about SGW selection using
UUT and NR Capability, see Dedicated Core Network and 5G EPC.
For information about how to configure the SGW selection based on IMSI
number series and Geographical Area, see Configure SGW Selection Based on
IMSI Number Series and Geographical Area for the MME.
The MME selects an SGW from the static SGW selection table when the following
configurations are performed:
The MME uses the Tracking Area Identifier (TAI) as the input to find candidate
SGWs from the static SGW selection table. The MME selects the SGWs according
to the priorities or capacities configured in the static SGW selection table. If the
priority and capacity of an SGW are both configured, the MME selects the SGW
by the priority, and then by the capacity. For multiple SGWs have the same
priority, the MME selects the SGW randomly.
When the Dedicated Core Network feature and Gateway Selection for NR Usage
feature, are enabled, the MME uses UUT and NR Capability as the extra filter to
select SGW from the candidate SGWs For more information about SGW selection
using UUT and NR Capability, see Dedicated Core Network and 5G EPC.
If the standalone PDN connectivity for the SGW candidates is already selected, at
additional PDN connections the current SGW is used as the only candidate.
The MME continues the gateway selection by using internal FQDNs based on the
configured SGW hostnames to sort the SGW candidates, see Sorting SGW
Candidates on page 27.
If the standalone PDN connectivity for the SGW candidates is already selected, at
additional PDN connections the current SGW is used as the only candidate.
The MME continues by sorting the SGW candidates, see Sorting SGW Candidates
on page 27.
The DNS query result is then filtered with an applicable subset of the services
required for the PDN connection. The services used depend on the applicable
interface and protocol, and are listed as follows:
— x-s5-gtp
— x-s5-pmip
— x-s8-gtp
— x-s8-pmip
When one, or both, of the two features Dedicated Core Network and Gateway
Selection for NR Usage are activated, multiple DNS S-NAPTR procedures are
performed with additional filtering.
When the Gateway Selection for NR Usage feature is activated and the UE is
allowed to use NR as secondary RAT, the following two S-NAPTR procedures are
performed.
The SGW candidates are sorted as described in section 4.12.1. For more
information about the Gateway Selection for NR Usage feature, see 5G EPC.
When the Dedicated Core Network feature is activated, the following two S-
NAPTR procedures are performed.
The SGW candidates are sorted as described in section 4.12.2. For more
information about the Dedicated Core Network feature, see Dedicated Core
Network.
Note: When both features are activated and the UE is allowed to use NR as a
secondary RAT, an additional S-NAPTR procedure is performed. This
procedure retrieves the SGW candidates supporting NR which are
dedicated to the specific UE Usage Type, for example, the filtering
candidates include both +nc-nr and +ue-N appended to the protocol.
These candidates have the highest priority when the PGW-SGW pairs
are sorted.
The services to be used are determined based on the protocol (GTP or PMIP) and
the interfaces used in the existing PDN connections. The interfaces that the SGW
candidates are required to support are determined based on the following cases:
The MME prefers an SGW supporting both S5 and S8 in the following cases:
— For a home subscriber with at least one S8 PDN connection.
— For a roaming subscriber with at least one S8 PDN connection, or when the
node function sgw_selection_based_on_all_apn_lbo is not active.
If there is no SGW supporting S5 and S8 available, the MME prefers the following
cases:
— For home subscribers, the MME finds an SGW supporting S5 interface.
— For roaming subscribers, if the majority of the existing PDN connections are
using S8, the MME finds an SGW supporting S8 interface; otherwise, finds an
SGW supporting S5 interface.
For more information about the S-NAPTR procedure, see DNS Description.
The sorting algorithm in the MME sorts the gateway candidates according the
following:
— If a configured static SGW exists, the MME sorts the candidate gateways
according to the gateway capacities and gateway priorities.
— In SGW selection based on the DNS query with TAI, the MME sorts the
candidates from the S-NAPTR procedure using both NAPTR records and SRV
records. NAPTR and SRV records contain information about the gateway
addresses and their relative weight in the network. The relative weight is
related to the load.
For information about SRV and NAPTR records sorting, see Sorting PGW
Candidates on page 20.
3.12.1 Optional Function for Sorting the SGW with the GTP-C Load Information
The Gateway Selection with GTP-C Load Information function enables the MME
to select the SGW based on the GTP-C LCI. This function helps to achieve a
balanced load network by using the dynamic load information provided in the
Load Control Information IE over the S11 interface.
Note: This feature is supported for home subscribers and roaming subscribers
whose subscription data and local configuration allow them to use local
breakout for all APNs. For roaming users with home routed traffic, the
GTP-C Network Load Control function is not applied.
This feature is activated based on the local configuration per PLMN. For
information about the configuration of gateway selection based on the LCI, see
Configuring Session and Mobility Management.
When the Gateway Selection with GTP-C Load Information function is activated,
the MME takes into account the node level LCI received from the SGW in addition
to the weight factors received in the DNS S-NAPTR procedure to select an
appropriate SGW.
To ensure an efficient load control in the network, the MME updates the load
control information if the new received load metric has an increase or decrease
During the gateway selection procedure, the MME computes the dynamic weight
factor for the candidate gateways within the same order and priority in the SRV
records. The MME uses the dynamic weight factor instead of the weight factor
received from the DNS server to sort the NAPTR and SRV records results in the
load sharing among the supported SGWs. Considering the current available load
and DNS weight-factor of the target node as follows.
For the NAPTR procedure, the dynamic weight factor is calculated using the
following formula:
For the SRV procedure, the dynamic weight factor is calculated using the
following formula:
Note: If the load metric is not available, the MME considers the load metric as
0 for a candidate target node.
For an example of the dynamic weight factor for each candidate SGW, refer to
Optional Function for Sorting the PGW with the GTP-C Load Information on
page 21.
The MME supports the LCI IE in the below messages sent from the SGW:
PGW candidate lists. For more information about sorted SGW candidate list and
sorted PGW candidate list, see Sorting SGW Candidates on page 27 and Sorting
PGW Candidates on page 20.
The MME builds all the possible gateway pairs by selecting SGWs from the sorted
SGW candidate list as the first node and then selecting PGWs from the sorted
PGW candidate list.
The following gateway pairs list shows all the possible gateway pairs based on a
sorted SGW candidate list with the SGW1 and the SGW2, and a sorted PGW
candidate list with the PGW1 and the PGW2:
— SGW1+PGW1
— SGW1+PGW2
— SGW2+PGW1
— SGW2+PGW2
If this parameter is set to off, the MME sorts the gateway pairs by the following
criteria in sequence as specified by 3GPP:
1. Colocation
Colocated gateway pairs have the highest priority. Gateways with the same
canonical node name are colocated in the same SGSN-MME, for example:
• topon.int.gw10.east.op.com + topon.pg.gw10.east.op.com
The MME analyzes the gateway names, and colocated gateway pairs are
allocated the highest priority.
If there are no more colocated gateway pairs, the MME considers non-
colocated gateway pairs with the topon label. Gateway pairs with the topon
label, for example, topon.int.gw10.east.op.com +
topon.pg.gw14.east.op.com, indicate the presence of a topological structure.
Non-colocated gateway pairs with the topon label are sorted by topological
closeness, which is indicated by the number of matching labels in the suffix
of the canonical node name. A pair of SGW and PGW with more matching
labels in the suffix domain name is closer to each other than other gateway
pairs.
If two or more gateway pairs have the same SGW, they are arrayed in the
same order as the PGWs in the sorted PGW candidate list.
If there are no more non-colocated gateway pairs with the topon label, the
MME considers non-colocated gateway pairs with at least one topoff label.
If two or more gateway pairs have the same SGW, they are arrayed in the
same order as the PGWs in the sorted PGW candidate list.
Note: — For static gateway selection, if there are no more colocated gateway
pairs, the MME considers non-colocated gateway pairs based on the
configured gateway priority and capacity. If two or more gateway
pairs have the same priority after sorting by the SGW order derived
according to the gateway priority and capacity, the MME sorts the
gateway pairs by the PGW order derived according to the gateway
capacities. The MME preferentially selects the gateway pair whose
PGW has a higher capacity. For more information about how the
MME sorts gateway candidates, see [unresolved external reference].
1. SGW order derived from either the DNS query procedure or the gateway
priority and capacity parameter from static SGW selection table.
a. Colocation
If two or more gateway pairs have the same priority after sorting by the
SGW order, they are arrayed in the same order as the PGWs in the sorted
PGW candidate list.
Note: — For static gateway selection, if there are no more colocated gateway
pairs, the MME considers non-colocated gateway pairs based on the
configured gateway priority and capacity. If two or more gateway
pairs have the same priority after sorting by the SGW order derived
according to the gateway priority and capacity, the MME sorts the
gateway pairs by the PGW order derived according to the gateway
capacities. The MME preferentially selects the gateway pair whose
PGW has a higher capacity. For more information about how the
MME sorts gateway candidates, see [unresolved external reference].
1. topoff.int.gw1.east.op.com
2. topon.int.gw2.east.op.com
3. topon.int.gw3.west.op.com
4. topoff.int.gw4.west.op.com
1. topon.pg.gw1.east.op.com
2. topon.pg.gw2.east.op.com
1. topoff.int.gw1.east.op.com + topon.pg.gw1.east.op.com
2. topoff.int.gw1.east.op.com + topon.pg.gw2.east.op.com
3. topon.int.gw2.east.op.com + topon.pg.gw1.east.op.com
4. topon.int.gw2.east.op.com + topon.pg.gw2.east.op.com
5. topon.int.gw3.east.op.com + topon.pg.gw1.east.op.com
6. topon.int.gw3.east.op.com + topon.pg.gw2.east.op.com
7. topoff.int.gw4.east.op.com + topon.pg.gw1.east.op.com
8. topoff.int.gw4.east.op.com + topon.pg.gw2.east.op.com
2. Among the non-colocated gateways, gateway pairs with the topon label are
prioritized over gateway pairs with at least one topoff label.
4. DNS priority is then used to determine the order among gateway pairs with
the same priority. For example, among the gateway pairs in steps 4 and 5,
and the gateway pairs in steps 6, 7, and 8 in the list of gateway pairs in the
following list.
1. topon.int.gw14.east.op.com
2. topon.int.gw10.east.op.com
1. topon.pg.gw19.west.op.com
2. topon.pg.gw10.east.op.com
1. topon.int.gw14.east.op.com + topon.pg.gw19.west.op.com
2. topon.int.gw14.east.op.com + topon.pg.gw10.east.op.com
3. topon.int.gw10.east.op.com + topon.pg.gw19.west.op.com
4. topon.int.gw10.east.op.com + topon.pg.gw10.east.op.com
4. topon.int.gw10.east.op.com + topon.pg.gw19.west.op.com
4. topon.int.gw10.east.op.com + topon.pg.gw19.west.op.com
hostname related to the S11 interface is not provided in the response, the SGSN-
MME performs an S-NAPTR procedure using the canonical node name of the
SGW and the following filter:
— x-3gpp-sgw:x-s11
The MME uses the FQDNs of the gateway pair to retrieve IP for the SGW and the
PGW.
To retrieve the SGW IP, if only IPv4 or IPv6 is configured for the S11 interface,
the MME sends a type A or a type AAAA query to the DNS. If both IPv4 and IPv6
are configured for the S11 interface, the MME sends both type A and type AAAA
queries to the DNS.
If the MME receives both an IPv4 and an IPv6 address for the SGW, the IPv6 is
prioritized over the IPv4 address. Thus if the MME supports IPv6, the MME uses
the received IPv6 address to set up a PDN connection.
To retrieve the PGW IP, the MME sends both type A and type AAAA queries to the
DNS.
If the MME receives both an IPv4 and an IPv6 address for the PGW, the MME
includes both addresses in the request for a PDN connection sent to the SGW.
If an SGW does not support the same IP version as the MME, all gateway pairs
that include that SGW are discarded from the list and the MME retrieves the IP
addresses of the next gateway pair in the list.
For an example and more detailed information about the DNS procedures to find
the SGW and the PGW, see DNS Description.
If a PDN connection setup fails because of the selected gateway pair, the MME
does the following to try to establish a PDN connection:
1. The MME discards the problematic gateway pairs from the gateway pair
candidate list as follows:
• If the MME does not receive a response from the SGW, the MME discards
the gateway pairs containing that SGW from the candidate list and
proceeds with .
• If the MME receives a rejection from the SGW or the PGW, the MME
discards the gateway pairs containing the rejecting gateway from the
candidate list.
The SGW or the PGW rejects a PDN connection setup attempt by the
MME with a GTPV2 cause code indicating a particular node-related
problem. The MME only proceeds with when the PDN connection setup
attempt is rejected with one of the GTPV2 cause codes configured
through the GwPairFallbackCauseList parameter. The following cause
codes are allowed for this parameter:
— 72: System Failure
2. The MME selects the next gateway pair in the gateway pair candidate list.
3. The MME tries to establish a PDN connection with the newly selected
gateway pair.
Note: The MME limits the fallback attempts to the value specified by the
GwMaxAttemptNumber parameter minus 1 . When the limit is reached,
the MME considers the PDN connection setup to have failed.
During mobility procedures, the SGW selection procedure can be triggered to find
the candidate SGW list. If the old SGW is included in this candidate list, the old
SGW continues to serve the UE. A new SGW is selected if the old SGW is not able
to serve the UE. During mobility procedures, the PGW is not changed.
— TAU procedures. For more information, see LTE Mobility Management and
Inter-System Mobility Management.
— S1- and X2-Based Handover procedures. For more information, see LTE
Mobility Management.
— Handover from non-3GPP networks to LTE. For more information, see Inter-
System Mobility Management.
Note: SGW selection is also performed during the PDN connectivity procedure
for UEs in EMM-REGISTERED without PDN Connection.
The SGW selection or reselection procedure for finding the SGW candidates is
performed as follows:
2. If static SGW selection is configured, the MME selects an SGW from the SGW
selection table. For more information, see Retrieving SGWs from Static SGW
Selection Table on page 24.
If the configured primary or secondary SGW is the same as the old SGW, the
same SGW is selected. If neither the primary SGW nor the secondary SGW is the
same as the old SGW, the SGW is changed to the primary SGW, or to the
secondary SGW if this SGW is configured when the primary SGW cannot be
reached.
Note: For a wildcard APN, the MME fetches the PGW identity from the
Specific-APN-Info IE. For other APNs, the MME fetches the PGW
identity from the MIP6-Agent-Info IE.
Table 1 shows how the MME selects a PGW based on the PGW identity. For more
information, see PGW and SGW Selection during PDN Connectivity Procedures
on page 12.
1. If there is an EMC PDN connection, the SGW with the highest priority toward
EMC PDN PGW is selected.
2. If no EMC PDN connection exists, or all SGWs have the same priority toward
the EMC PDN PGW and an IMS PDN connection exists, the SGW with the
highest priority toward the IMS PDN PGW is selected.
3. If no EMC PDN connection or IMS PDN connection exists, or all SGWs have
the same priority toward the EMC PDN PGWs and IMS PDN PGWs
respectively, then the SGWs are prioritized toward all non-EMC and non-IMS
PDN PGWs in the following manner:
a. If there are SGWs that are colocated with non-EMC PDN PGWs and
non-IMS PDN PGWs, the colocated SGWs are prioritized by the number
of PDN connections supported by the PGW. The SGW colocated with the
PGW supporting the most PDN connections is prioritized over other
SGWs.
b. If there are non-colocated SGWs for which topology applies, the SGWs
are prioritized by their topological closeness to the PGWs.
c. If there are gateway pairs that are non-colocated, or if topology does
not apply, or some SGWs still have the same priority after colocation and
topological sorting, the SGW order derived from the DNS query
procedure is used to sort the SGWs.
— Inter-MME TAU
— Inter-MME Handover
— MME moves
PGW reselection can be enabled using the pgw_reselection parameter for the
PDNs configured in the APN list. For information about how to configure the
PGW Reselection function, see Configuring Session and Mobility Management.
For each PDN in the specified APNs, the new MME compares the PGW used in
the old area with the PGW candidates list result of the Gateway Selection
function in the new MME area based on the FQDN or the IP address. The
comparison type depends on the information that the new MME receives from
the old Gn/Gp-SGSN or the old MME, and on the information that the new MME
receives based on the Gateway Selection function.
— If the PGW used in the old area does not match any member of the PGW
candidates list, PGW reselection is performed. The PGW used in the old area
for the PDN is indicated by its FQDN or IP address. For information about the
Gateway Selection function, see PGW and SGW Selection during PDN
Connectivity Procedures on page 12.
— If the PGW used in the old area matches one member of the PGW candidates
list, and there are PGWs with different priorities created according to the
configuration of one of the following rules, the MME also checks the
pgw_reselection_for_prioritized_one node function to decide whether
to reselect a prioritized PGW:
— The SMF rule is configured as smf_prioritized or pgw_prioritized.
If the node function is set to its default value on, the MME reselects a
prioritized PGW. If the node function is set to off, the PDN established in the
less prioritized PGW is not deactivated. This triggers fewer PDN
deactivations and activations, and minimizes session interruptions.
If the Gateway Selection for NR Usage feature is enabled, the new MME takes
into account the NR network capability when constructing the PGW candidates
list. For example. based on the NR rule configuration, the new MME can reselect
an NR-capable PGW when the PGW in use does not support NR and there is at
least one NR-capable PGW in the PGW candidates list. For more information, see
5G EPC.
If the Dedicated Core Network feature is enabled, the new MME takes into
account the UUT rule when constructing the PGW candidates list. For example,
based on the UUT rule configuration, the new MME can reselect a PGW that
supports the currently used UUT of the UE. For more information, see Dedicated
Core Network.
If the Advanced 4G/5G Interworking feature is enabled, the new MME takes into
account the SMF network capability when constructing the PGW candidates list.
For example, based on the SMF rule configuration, the new MME can reselect a
combined PGW-C and SMF. For more information, see Advanced 4G/5G
Interworking.
If PGW reselection is needed, the new MME performs the following actions:
The MME then reselects a PGW for the PDNs when the UE reactivates the PDN
connections or reattaches to the network.
6 PDN Disconnection
When the MME initiates a PDN disconnection from the last PDN connection, the
MME-initiated Detach procedure is performed if the UE does not support EMM-
REGISTERED without PDN Connection. For more information about this, see LTE
Mobility Management.
2. The MME sends a Delete Session Request message to the PGW through
the SGW. This message contains information about the default bearer that
belongs to the PDN connection that is to be deleted.
If the NR Usage Data Reporting feature is activated and if the the UE is NR-
capable and not restricted from using NR, step 2 is executed after step 4 to
allow the MME to include the reports described in step 4.4 in the Delete
Session Request message.
1. The MME initiates the deactivation of all bearers associated with the PDN
connection to the eNodeB by sending the E-RAB Release Command
message to the eNodeB. This message contains the Deactivate EPS
Bearer Context Request message for the default bearer. In this step,
the MME recalculates the UE-AMBR and provides it in the message to the
eNodeB. For more information about how the AMBR is calculated, see
AMBR on page 3.
1. The MME sends a Downlink NAS Transport message to the UE. The
Downlink NAS Transport message contains the Deactivate EPS
Bearer Context Request message.
last one. When the last PDN connection is to be removed, the MME-Initiated
Detach procedure is executed if the UE does not support EMM-REGISTERED
without PDN Connection. For more information, see LTE Mobility Management.
1. The PDN Disconnection procedure is initiated by the PCRF through the PGW.
The PGW sends the Delete Bearer Request message through the SGW to
the MME to deactivate the selected PDN connection identified by the LBI.
When the last PDN connection that belongs to a UE is released, and the UE
does not support EMM-REGISTERED without PDN Connection, the MME
starts an MME-Initiated Detach procedure without signaling the SGW. For
more information, see LTE Mobility Management.
If the Delete Bearer Request message from the PGW does not include
reactivation requested, and the UE is in the ECM-IDLE state, and are
skipped.
a. The MME initiates the deactivation of all bearers associated with the
PDN connection to the eNodeB by sending the E-RAB Release
Command message to the eNodeB. This message contains the
Deactivate EPS Bearer Context Request message for the
default bearer. In this step, the MME recalculates the UE-AMBR and
provides it in the message to the eNodeB. For more information
about how the AMBR is calculated, see AMBR on page 3.
a. The MME sends a Downlink NAS Transport message to the UE. The
Downlink NAS Transport message contains the Deactivate EPS
Bearer Context Request message.
4. The MME deletes the bearer context related to the deactivated EPS bearer,
and acknowledges the bearer deactivation to the PGW by sending a Delete
Bearer Response message containing the default Bearer Identity. The
message is sent through the SGW to the PGW.
The MME can include the Secondary RAT usage data report IE, if
received in step 3.4.
The HSS-Based P-CSCF Restoration function is applicable for UEs with an IMS
PDN connection. For information about IMS PDN connections, see IMS APN
Identification on page 4.
When the MME receives the Insert Subscriber Data Request message from
the HSS indicating P-CSCF Restoration, the MME checks if the HSS-based P-
CSCF Restoration function is enabled. The subsequent MME behavior is based on
the configuration of the HssBasedPcscfRestoration parameter.
— If the UE is in the ECM-IDLE state, the MME pages the UE and performs the
preceding operation when the UE becomes ECM_CONNECTED.
If the MME fails to notify the UE of the IMS PDN deactivation with
reactivation requested, the MME keeps the IMS PDN instead of deleting
it. The kept IMS PDN is deactivated with reactivation requested at the
next UE activity, or the next P-CSCF Restoration with successful paging. As a
result, the PDN disconnection or detach is triggered.
When the UE reactivates the IMS PDN or reattaches to the network, the UE
selects a new P-CSCF.
Note: If the UE is in the ECM-IDLE state, the MME first pages the UE. In this
case, the P-CSCF Restoration indication is sent in the Modify Bearer
Request message during the Paging Response procedure.
When the UE receives the Modify EPS Bearer Context Request message
from the MME with a list of available P-CSCF addresses in the PCO IE, the
UE selects an available P-CSCF from the list and performs a new initial IMS
registration.
After the IMS PDN is released, the UE selects a new P-CSCF when the UE
reactivates the IMS PDN or reattaches to the network.
7.2.1 Limitation
The optional extension function requires support in the S-CSCF and the HSS and
has the following additional limitations:
2. If the UE is in the ECM-IDLE state, the MME initiates the Paging procedure
which is followed by the Service Request procedure. These procedures are
further described in LTE Mobility Management.
3. The MME selects an EPS Bearer Identity and sends an E-RAB Setup
Request message containing the EPS Bearer Identity to the eNodeB. The E-
RAB Setup Request message contains one Activate Dedicated EPS
Bearer Context Request message for each bearer to be set up.
The following flows only describe the parts related to dedicated bearer
activation. The complete Attach procedure flow is described in LTE Mobility
Management. The complete PDN Connectivity procedure flow is described in
Traffic Case on page 7.
1. The SGW sends the Create Bearer Request message piggybacked to the
Create Session Response message, after initiation from the PGW.
2. The MME sends the Initial Context Setup Request message, including
the Activate Default EPS Bearer Context Request message and the
Activate Dedicated EPS Bearer Context Request message, to the
eNodeB which forwards it to the UE.
3. The eNodeB sends the Initial Context Setup Response message to the
MME, as in the normal Attach procedure.
6. The MME sends the Create Bearer Response message with the Modify
Bearer Request message piggybacked to it.
7. The SGW sends the Create Bearer Response message to the PGW.
8. The SGW sends the Modify Bearer Response message to the MME.
Dedicated bearers are activated and deactivated depending on the service used.
The network decides whether dedicated bearers need to be deactivated.
Note: If the dedicated bearer is with QCI =1, and the PDN connection is
marked as to be reactivated because of PGW reselection, the MME
initiates the PDN deactivation with the cause code reactivation
requested, instead of deactivating the dedicated bearer.
If the Delete Bearer Request message from the PGW does not include
reactivation requested, and the UE is in the ECM-IDLE state, to are
skipped.
5. The UE releases the bearers and acknowledges this by sending the RRC
Connection Reconfiguration Complete message to the eNodeB.
8. The MME deletes the bearer context related to the deactivated EPS bearer,
and acknowledges the bearer deactivation by sending a Delete Bearer
Response message containing the EPS bearer identities to the PGW through
the SGW.
The MME can include the Secondary RAT usage data report IE, if
received in step 6.
— The target eNodeB could not set up all dedicated bearers during the S1-
Based Handover procedure.
— The target eNodeB has not setup all dedicated bearers during the X2-Based
Handover procedure.
— The eNodeB could not set up all dedicated bearers during the Service
Request procedure.
— A failure message is received from the SGW during the Service Request
procedure.
— A bearer mismatch occurs in the EPS Bearer Status IE in the TAU procedure.
— The conversion between the PDP contexts and the EPS Bearers fail.
— If the active EPS GBR bearers with 0 MBR are received from an old SGSN in
an inter-system change procedure.
The Handover, TAU, and Service Request procedures are described in LTE
Mobility Management.
The mapping between PDP contexts and EPS Bearers is described in Inter-
System Mobility Management.
When the timer expires, the MME handles the GBR bearers according to the
following:
— If there is no ongoing paging for the UE, the MME initiates deactivation of
the GBR bearers.
— If there is ongoing paging for the UE, the MME keeps the GBR bearers if the
paging is successful, and deletes the GBR bearers if the paging fails.
Any of the following S1AP cause codes can be configured for Timer-Based
Deactivation of GBR Bearers:
— RADIO_CONNECTION_WITH_UE_LOST
— All Radio Network Layer category-related cause codes except the normal
S1AP cause codes that are described in the following list
— All cause codes except the normal S1AP cause codes that are described in
the following list
The following five cause codes are considered as normal S1AP cause codes, and
are not applicable for Timer-Based Deactivation of GBR Bearers:
— USER_INACTIVITY
— INTER_RAT_REDIRECTION
— CS_FALLBACK_TRIGGERED
— UE_NOT_AVAILABLE_FOR_PS_SERVICE
— REDIRECTION_TOWARDS_1XRTT
UEs reconnecting to the network within the specified time period with preserved
bearers. For more information about the toolbox script, see Toolbox Description.
When receiving the Downlink Data Notification message during the running
of the timer, the MME behaves differently.
• If all the bearers indicated by the EPS bearer identities in the Downlink
Data Notification message are dedicated bearers or the message
does not contain any EPS bearer identity, the MME keeps the timer
running and does not start the Paging procedure.
1. E-RAB Release
Indication
2. Delete Bearer
Command
3. Delete Bearer
Command
4. Delete Bearer
Request
5. Delete Bearer
Request
6. Delete Bearer
Response
7. Delete Bearer
Response
compulsory communication
conditional or optional communication
2. The MME sends a Delete Bearer Command message per PDN connection to
the SGW to deactivate the selected dedicated bearers.
3. The SGW sends a Delete Bearer Command message per PDN connection to
the PGW.
5. The SGW forwards this Delete Bearer Request message to the MME.
6. The MME deletes the bearer context related to the deactivated PDN
connection, and acknowledges the deactivation by sending a Delete
Bearer Response message containing the EPS bearer identities to the PGW
through the SGW.
7. The SGW deletes the bearer context related to the deactivated PDN
connection, and acknowledges the deactivation by sending a Delete
Bearer Response message containing the EPS bearer identities to the PGW.
4. Delete Bearer
Request
5. Delete Bearer
Request
6. Delete Bearer
Response
7. Delete Bearer
Response
compulsory communication
conditional or optional communication
2. The MME sends a Delete Bearer Command message per PDN connection to
the SGW to deactivate the selected dedicated bearers.
The Secondary RAT usage data IE is included if received from the eNodeB.
3. The SGW sends a Delete Bearer Command message per PDN connection to
the PGW. The SGW includes the Secondary RAT usage data IE in this
message if it was received in the previous message and if the PGW is the
intended receiver.
5. The SGW forwards this Delete Bearer Request message to the MME.
6. The MME deletes the bearer context related to the deactivated PDN
connection, and acknowledges the deactivation by sending a Delete
Bearer Response message containing the EPS bearer identities to the PGW
through the SGW.
7. The SGW deletes the bearer context related to the deactivated PDN
connection, and acknowledges the deactivation by sending a Delete
Bearer Response message containing the EPS bearer identities to the PGW.
10 Modification of Bearers
The network decides whether dedicated bearers need to be modified. Since the
APN-AMBR can be changed because of the PGW-initiated Bearer Modification
procedure, the UE-AMBR can also need to be changed. The UE-AMBR is set by
the MME to the sum of the APN-AMBR of all active APNs up to the value of the
subscribed UE-AMBR.
Figure 11 shows the two possible traffic cases to follow depending on whether
the UE is in the ECM-IDLE or the ECM-CONNECTED state.
5. The MME sends a Modify EPS Bearer Context Request message, which
contains the EPS Bearer ID, the TFT, and the APN-AMBR, to the UE through
the eNodeB. This step is repeated for each bearer to be modified.
6. The UE responds to the MME with a Modify EPS Bearer Context Accept
message, which contains the EPS Bearer ID, through the eNodeB to the
MME. This step is repeated for each bearer to be modified.
7. Upon reception of the Modify EPS Bearer Context Accept message, the
MME acknowledges the bearer modification by sending the PGW an Update
Bearer Response message containing the EPS Bearer ID through the SGW.
The PGW can also inform the PCRF of the bearer modification.
7. Direct Transfer /
Modify EPS Bearer Context Accept
8. Uplink NAS Transport /
Modify EPS Bearer Context Accept
Note: If the UE is in the ECM-IDLE state and only the ARP is modified in
the Update Bearer Request message, the UE and eNodeB are not be
notified, the MME sends the Update Bearer Response message to
the SGW directly.
2. If the UE is in the ECM-IDLE state, the MME initiates the Paging procedure
followed by the Service Request procedure, where the MME uses the old QoS.
These procedures are further described in LTE Mobility Management.
3. The MME sends the E-RAB Modify Request message to the eNodeB
containing the EPS Bearer Identity, the bearer level QoS, Modify EPS
Bearer Context Request message, and the UE-AMBR, if it is changed. The
Modify EPS Bearer Context Request message contains the EPS Bearer
Identity, the bearer level QoS, the APN-AMBR, and the TFT.
4. The eNodeB maps the modified bearer level QoS to the Radio Bearer QoS. It
sends an RRC Connection Reconfiguration message containing the
Radio Bearer QoS, the Modify EPS Bearer Context Request, and the EPS
Bearer Identity to the UE.
7. The UE sends a Direct Transfer message containing the Modify EPS Bearer
Context Accept message to the eNodeB. The Modify EPS Bearer
Context Accept message includes the EPS Bearer Identity.
9. Upon reception of the E-RAB Modify Response message in and the Uplink
NAS Transport message in , the MME acknowledges the bearer modification
to the SGW by sending an Update Bearer Response message. The SGW
forwards this message to the PGW that informs the PCRF.
2. The MME stores the updated subscription data and sends an Insert
Subscriber Data Answer message to the HSS. If the subscribed UE-AMBR
is modified, the MME calculates a new UE-AMBR value.
If the UE is in the ECM-IDLE state, the new UE-AMBR is stored for later use.
Note: If the subscribed RFSP is changed for the home subscriber, the new
RFSP is sent to the eNodeB within the same UE Context
Modification Request message.
If the subscribed QoS profile changes in the MME when the UE is suspended
during the CSFB procedure, the HSS-initiated Subscribed APN-AMBR or Bearer
Level QoS Modification procedure is performed after the UE returns to the E-
UTRAN and resumes.
1. If the APN-AMBR or bearer level QoS in the HSS is modified, the HSS sends
an Insert Subscriber Data Request message to the MME. This message
contains the modified APN-AMBR or bearer level QoS.
2. The MME sends an Insert Subscriber Data Answer message to the HSS.
3. The MME sends a Modify Bearer Command message for each PDN
connection to the PGW through the SGW. This message includes the APN-
AMBR and bearer level QoS.
4. The PGW sends the Update Bearer Request message containing the APN-
AMBR or bearer level QoS to the MME through the SGW.
5. If the UE is in the ECM-IDLE state, the MME initiates the Paging procedure
followed by the Service Request procedure which is further described in LTE
Mobility Management.
6. The MME continues the bearer modification with the QoS update procedure,
which is described in Traffic Case on page 66.
The IRAT TAU from GSM or WCDMA to LTE and IRAT Handover from WCDMA to
LTE procedures are described in Inter-System Mobility Management.
Additionally, the QoS modification can also be triggered by inter-MME TAU and
inter-MME handover from an MME not supporting NR to an MME supporting NR
if all of the following conditions are met:
— The length of the Extended Access Restriction is not received from the old
MME and the value of the APN-AMBR in use is less than 4.29G.
TAU Procedure
The following steps describe the QoS Modification in IRAT Mobility procedure:
1. The MME sends a Modify Bearer Command message for each PDN
connection to the PGW through the SGW. This message includes the updated
QoS information.
2. The PGW sends the Update Bearer Request message containing the APN-
AMBR and bearer level QoS to the MME through the SGW.
3. If the UE is in the ECM-IDLE state, the MME initiates the Paging procedure
followed by the Service Request procedure which is further described in LTE
Mobility Management.
4. The MME continues the bearer modification with QoS update procedure,
which is described in Traffic Case on page 66.
If the UE and the MME use DoNAS, the MME establishes a GTP-U tunnel to
transfer the user data between the MME and the SGW over the S11-U interface.
However, if the S11USupportInGW parameter is set to off, the MME tunnels the
user data over the S11-U interface in an S1-U eNodeB F-TEID IE instead. The
MME then includes an S1-GTP-U F-TEID in the Modify Bearer Request
message sent to the SGW.
The SGW acknowledges the request by including an SGW GTP-U F-TEID either
in the Create Session Response or in the Modify Bearer Response message
sent to the MME.
compulsory communication
optional communication
3. After receiving the Bearer Resource Command message from the MME, the
SGW sends a Bearer Resource Command message to the PGW.
For Procedure 2, see the section PGW-Initiated Bearer Modification with Bearer
QoS Update on page 66.
Note: — This procedure only supports the Bearer Resource Modification for
S1-U PDN connection.
The MME can include a RAN/NAS cause code in the RAN/NAS Release Cause IE
in the messages sent to the SGW if the following conditions are fulfilled:
Table 8 lists the messages and the scenarios where the MME can forward the
RAN/NAS cause code to the SGW.
Table 8 Messages and Scenarios Where RAN or NAS Cause Codes Are
Forwarded to the SGW
Messages Scenarios
Delete Session Request MME-Initiated Detach
HSS-Initiated Detach
MME-Initiated PDN Disconnection
Delete Bearer Command MME-Initiated Dedicated Bearer
Deactivation(1)
Delete Bearer Response The MME receives a PGW-initiated
Bearer Deactivation Request while
waiting for GBR deactivation triggered
by an S1 Release.
Create Bearer Response A failing PGW-initiated Dedicated
Bearer Activation
Update Bearer Response A failing PGW-initiated Bearer
Modification
(1) When the S1TimerGbrBearerDeactivation timer is running in an S1 release, an interfering
procedure stops the S1 release and the S1TimerGbrBearerDeactivation timer. If the interfering
procedure triggers a Paging procedure and the paging fails because of time-out, the GBR
Bearer Deletion procedure is triggered and the S1AP cause code for the S1 release is forwarded
in the GBR Bearer Deletion procedure.
13.1 Parameters
To display the configuration classes and parameters related to LTE Session
Management, use the get_config_area CLI command for the following
configuration areas:
— GwFailureRestoration
— GwBlacklisting
— ImsiBasedGwSelection
— InterfaceS11
— InterfaceS1MME
— PgwReselection
— QualityOfService
— StaticGwSelection
— StaticSgsnMmeSelection
13.2 Counters
The following counters and PmGroups are valid for LTE Session Management:
— SGSN-MME_GTP_Interface
— SGSN-MME_GTPControlPlane
— SGSN-MME_GTPUserPlane
— SGSN-MME_GTPUserPlane_E
— SGSN-MME_Session_DiscardedMessages_G
— SGSN-MME_Session_HO
— SGSN-MME_Session_SM_E
— SGSN-MME_Session_SM_E_ARP
— SGSN-MME_Session_SM_E_QCI
— gtpMMEPathFailure
— gtpMscPathFailure
— gtpMscPathFailure
— pgwRestarted
— sgwRestarted
13.4 Logs
The Session Event Log is valid for LTE Session Management.
For descriptions of the log file, refer to Operation and Maintenance Description
and Session Event Log.
14 Compliance