Networks vMME Operator Guide
Networks vMME Operator Guide
Abstract: Affirmed Networks vMME Operator Guide for MME8.1 including MME and SGSN functionality
Copyright© 2015 Affirmed Networks, Inc. All Rights Reserved.
This document contains information that is the property of Affirmed Networks, Inc. This document may not
be copied, reproduced, reduced to any electronic medium or machine readable form, or otherwise
duplicated, and the information herein may not be used, disseminated or otherwise disclosed, except with
the prior written consent of Affirmed Networks, Inc.
References
11 Glossary......................................................................................................................................... 157
Configuration procedure samples for a standalone MME, and a standalone SGSN, and variations
on a combined node.
Pre-Requisites:
This document assumes that the reader has a good understanding of 3GPP networks and is
familiar with tailf Command Line Interface (CLI).
vMME Operations Administration and Maintenance (OAM) details – please refer to the OAM
Reference Guides (for CM, CLI, FM, PM, CSL):
The Operator’s Guide is considered a layer above the reference guides and is more procedure
oriented.
The Monitoring Guide is intended for end customers to help diagnose systems using formally
modeled alarms, counters and logs. Issues that are not resolved by reading this should be
referred to support teams.
The following sections were added as a result of these release 8.1 features:
AGW-23400 Local Config for VLR SV address resolution – New section Sv to VLR
AGW-24479 GTP Timers IP Config S3-S10-S16-SV – New sections Sv GTP Mobility, S16 GTP
Mobility and Partial Nodal Reset for GTP
AGW-24205 Non-VLAN Support for Load Balancer – New section Non-VLAN Support for Load
Balancer
The following sections were updated as a result of these release 8.1 features:
AGW-23400 Local Config for VLR SV address resolution – Updated sections SGs to VLR and Gs
– VLR (MSC)
AGW-24479 GTP Timers IP Config S3-S10-S16-SV – Updated section GTP Feature Overview
The following sections describe VNF (MME-SGSN virtual network function) configuration applied from
the CLI of the active MGMT VM.
[edit]
user@host% show system timezone
timezone America/Chicago;
[ok][2014-10-20 01:49:06]
Optionally within ‘system’, you can set contact and location information.
user@host % set system location "Ricahrdson, TX. Lab 2, Rack 3, Shelf 4"
[edit]
user@host% show ntp
server nist.time.nosc.us {
ip-address 96.226.242.9;
}
[ok][2014-10-20 01:50:56]
The MME supports periodic PM output in 3GPP XML format at three different intervals of 5, 15, or 30
minutes. The PM XML output is in the following directory on the MGMT VM:
/usr/local/epc/statistics/reports.
Step # Action
1 Configure the collection interval and retention policy.
In this example we will collect every 15 minutes, and retain the maximum reports allowed.
user@host> configure
Entering configuration mode private
[ok][2014-05-07 21:25:43]
[edit]
user@host% set mme-sgsn statistics collection-interval 15
user@host% set mme-sgsn statistics retention-report 288
user@host% commit
Commit complete.
[ok][2014-05-07 21:59:38]
Note: The retention limits vary depending on the collection interval configured.
2 Check the configuration.
The floating IP address for the NMS interface to the active MGMT VM.
Step # Action
1 oam ip interface configuration:
user@host> configure
Entering configuration mode private
[ok][2014-05-07 21:25:43]
[edit]
user@host% set mme-sgsn oam-ip-address 1.2.3.4
user@host% commit
Commit complete.
[ok][2014-05-07 21:59:38]
Note:
A complete list of all service parameters supported by the MME/SGSN and the
associated procedures and adjacent nodes to be selected is provided in the DNS
section of the vMME Product Overview document.
Caution:
Review all entries in the static cache or on the DNS server carefully and check
compliancy to applicable standards and MME/SGSN documentation mentioned
above. Any errors, even typographical, can lead to call processing failures.
WARNING:
In order to ensure both static and external DNS queries complete successfully, the
DNS I/F must be configured.
Note: In configurations where an external DNS is not available or present, the
DNS interface must still be configured.
Example: set mme-sgsn dns dns-if ip-address 10.0.0.0
For configurations using an external DNS server, please also refer to the DNS vendor’s user guide.
Step # Action
1 Define local DNS Interface. A loopback interface is required for the DNS IP address.
user@host% set mme-sgsn dns dns-if dscp-mark de dynamic-cache-size 5 initial-retry-timer 500 ip-
address 172.18.248.14 max-retry-timer 10 negative-cache-ttl 300 override-ttl off retry-count 6
routing-instance EPCCore01
3.1.3.1 A-Record
Step # Action
1 Define local A-Record Replacement.
3.1.3.2 NAPTR-Record
Step # Action
1 Define APN-FQDN for PGW lookup.
Step # Action
apn.apn.epc.mnc011.mcc111.3gppnetwork.org services x-3gpp-pgw:x-s5-pmip:x-s5-gtp replacement
traffic-test order 1 preference 1 flags 'a'
domain-name test-apn.apn.epc.mnc011.mcc111.3gppnetwork.org;
services x-3gpp-pgw:x-s5-pmip:x-s5-gtp;
replacement traffic-test;
order 1;
preference 1;
flags 'a';
3.1.3.3 SRV-Record
Step # Action
1 Add srv-record into static cache.
user@host% set mme-sgsn dns static-cache srv-record 1 node-name sss port 2 priority 1 target-name
a1 weight 12
user@host% show mme-sgsn dns static-cache srv-record
1 {
node-name sss;
target-name a1;
port 2;
priority 1;
weight 12;
}
}
3.1.3.4 AAAA-Record
Step # Action
1 Configure aaaa-record and display the details.
Syntax:
user@host% set mme-sgsn dns static-cache aaaa-record <entry-index> node-name <node value > node
node-ip <node-ip-address>
3 {
node-name ip.sgw2.rch.tx.affirmed-cta.com;
node-ip-address 2280::ec01:213:172:236:8;
}
SNMPv2 is used for reporting of alarms and logs to the NMS. For instructions to view logs and
manage alarms and logs, refer to the vMME Fault Management guide.
Step # Action
1 Configure the target to which traps are sent.
The following configuration enables the TCP socket in default routing-instance for SNMP master agent on the MGMT
VM.
user@host% set snmp trap-targets 172.16.239.47 community public
The following configuration enables the TCP socket for SNMP Managers
user@host% set snmp managers 172.16.239.50 community public
The MME/SGSN provides on-board continuous health monitoring capability on a per SC service
instance basis. It will periodically compute the most important KPI (attach and activation success rate,
as defined in the vMME Monitoring Guide), and raise an alarm if the configurable threshold values are
met. This capability is ‘on’ by default and the default values may be reasonable for some operators.
However, the following parameters are adjustable. This applies to GPRS, UMTS, and LTE
technologies. The corresponding alarms are:
The algorithm is based on a sliding window of samples, taken at every ‘update-period’ and averaged
over the ‘window’ duration. If ‘min-data-threshold’ samples are not taken for a given ‘update-period’,
then the KPI is not updated (this minimizes the risk of spurious alarms during low traffic periods).
Step # Action
1 Display the default health profile (default values shown).
[edit]
admin@host% show mme-sgsn status profile
update-period 3;
window 15;
min-data-threshold 100;
attach-major-threshold 90;
attach-minor-threshold 95;
attach-clear-threshold 98;
act-major-threshold 90;
act-minor-threshold 95;
act-clear-threshold 98;
[ok][2015-01-21 09:38:10]
[edit]
admin@host% set mme-sgsn status profile update-period 5
[edit]
admin@host%
The MME/SGSN collects performance monitoring statistics on a per service instance (ie. VM) basis
for the most part. There is an option to enable collection of selected statistics groupsets on a per peer
entity basis, but this is off by default due to the potentially large volume of data that it may produce.
Enabling these features would only be recommended for a short duration of critical analysis or on
small systems. For MME, this includes statistics per eNB and per FXA RADIUS peer, and for both
MME and SGSN per GTP path. These boolean controls are located under the feature control table.
Only those group-sets explicitly listed below are activated by these features. When the ‘Feature
Status’ value is ‘off’, the associated group-sets do not appear on the CLI or in the PM report system.
→ s1Enb
→ gtpV1Path
→ gtpV2Path
→ fxaPeer
Step # Action
1 Display the current feature control settings.
control direct-tunnel {
status on;
}
control fxa {
status on;
}
control gs {
status on;
}
control hspa-plus {
status on;
}
control lawful-intercept {
status on;
}
control s101 {
status on;
}
control s102 {
status on;
}
control s4-sgsn {
status on;
}
control sbc {
status on;
}
Step # Action
control sgs {
status on;
}
control trace {
status on;
}
2 Enable the stats features as needed.
Caution:
status on;
status on;
status on;
This feature augments the ‘sshd’ PAM service (on MME-SGSN OAM interfaces) by inserting
pam_tacacs into its auth, account and session stacks. Augmented ‘sshd’ PAM service stacks are
arranged to fall-back/fall-through to the default ‘sshd’ PAM service stack’s modules (mostly comprised
of common-[auth|account|session]) in the event that TACACS+ servers are unreachable. The
augmented account stack derives a “Group ID” (GID) from the service(s) users have configured in
TACACS+. The derived GID is the basis for the user’s access control.
Once MME-SGSNs have TACACS+ servers configured, users can be administrated centrally,
eliminating the need to manually create user accounts on each MME-SGSN node.
The MME-SGSN CLI is accessed via ssh to the (floating) OAM IP address jointly owned by the
management VMs. User accounts are created so that their shells launch the CLI.
The following procedure describes TACACS+ server configuration for “shrubbery” tac_plus. The
following requires sudo (or root) on the host running the tac_plus server. For the purpose of this
example, the IP address of the tac_plus server host is 192.168.40.86.
Step # Action
1 Configure tac_plus secret key in /etc/tacacs+/tac_plus.conf
1a Edit /etc/tacacs+/tac_plus.conf as root
sudo vi /etc/tacacs+/tac_plus.conf
key = testing123
sudo vi /etc/tacacs+/tac_plus.conf
user = bob {
global = cleartext bob
service = hcta-admin {}
}
The following procedure describes MME-SGSN configuration corresponding to the tac_plus server
configured in the preceding table.
Step # Action
1 Configure common TACACS+ authentication/authorization mandatory properties
[edit]
admin@host% set aaa tacacs-plus authentication auth-type login
[ok][2014-12-10 21:40:18]
[edit]
admin@host% set aaa tacacs-plus authentication server server1 ip-address 192.168.40.86 secret
testing123
[ok][2014-12-10 21:50:40]
[edit]
admin@host% set aaa tacacs-plus accounting server server1 ip-address 192.168.40.86 secret
testing123
[edit]
admin@host% commit
Commit complete.
[ok][2014-12-10 22:00:14]
Caution:
The CSL feature can generate large volumes of data depending on which events
are enabled and depending on the user traffic load on the system. Choose the
events and triggers carefully. By default all failure events are enabled.
CSL application is configured automatically to run on the MGMT VM (active and standby).
Step # Action
1 Configure CSL Events and File Behavior.
Step # Action
triggers-mme fail,reject,release;
max-records-per-period no-limit;
Configure to close the file every 5 minutes:
Login into the MGMT VM (determine the MGMT IP from the OpenStack UI)
%ssh mtc@172.24.255.253
Password:
mtc@vse01-vm01:~$cd /usr/local/epc/csl/closed
mtc@vse01-vm01:/usr/local/epc/csl/closed$ ls -l
total 4
-rw-r--r-- 1 root root 18 Dec 23 15:40 csl-MME-2-20131223154030-0600
mtc@vse01-vm01:/usr/local/epc/csl/closed$ cd ../open
mtc@vse01-vm01:/usr/local/epc/csl/open$ ls -l
total 4
-rw-r--r-- 1 root root 18 Dec 23 15:41 ocsl-MME-2-20131223154100-0600
mtc@vse01-vm01:/usr/local/epc/csl/open$
mtc@vse01-vm01:~$ csl-decode -h
Usage: /usr/local/epc/bin/csl-decode [<option>] [<filter>] <csl_file_list>
where <option> can be one of the following
-b will print hex format output (you can give -binary also)
-d will print detailed output (you can give -detailed also)
-brief print output in brief mode
-summary print output in summary mode
-csv print output into a file, in xls readable format(.csv)
where <filter> can be one of the following
-release print only release CSL records
-reject print only reject CSL records
-failure print only failure CSL records
-success print only success CSL records
-abnormal-release print only abnormal-release CSL records
-normal-release print only normal-release CSL records
and <csl_file_list> can be one or more csl files
Only files of name beginning with csl or ocsl are processed
Filters can be used only with -d and -brief option
Step # Action
When filters are used with out any option, -d is used by default.
mtc@vse01-vm01:/usr/local/epc/csl/closed$
mtc@vse01-vm01:/usr/local/epc/csl/closed$ csl-decode -d csl-MME-2-20131223154030-0600
CSL file transfer may be initiated by the NMS on a periodic basis. File deletion is to be done by the
NMS so it does not have to track what remains to be pulled.
Step # Action
1 SFTP to the MME/SGSN (non-gui sftp client is shown but any sftp client should work). Use the oam-ip-
address as provisioned on the MME/SGSN in section 3.1.2.2
sftp sftpuser@<sftpip>
sftp> cd /usr/local/epc/csl/closed
sftp> ls
csl-MME-4-20120419115100-0500 csl-MME-4-20120419120830-0500
4 Make a new directory to move the files to pull into (a name besides ‘pull’ can be used).
5 Move (rename) the closed files into the new ‘pull’ directory.
ALERT: Use the above command for each file to be pulled. A graphical sftp client will make this much easier since
multiple files can be highlighted and dropped into the ‘pull’ folder.
ALERT: A new directory is created because the MME/SGSN may still be writing new files to disk while the files are
being pulled. This allows for an easy way to tell which files should be deleted after the pulling is complete. Some
graphical clients allow you to ‘move’ files (copy and then delete) directly from remote to local and thus the need to
move them to a temporary directory is no longer necessary.
6 Get the files from the ‘pull’ directory.
sftp> cd pull
sftp> get *
Fetching /usr/local/epc/csl/pull/csl-MME-4-20120419115100-0500 to csl-MME-4-20120419115100-0500
Fetching /usr/local/epc/csl/pull/csl-MME-4-20120419120830-0500 to csl-MME-4-20120419120830-0500
sftp> lls –l
ALERT: The command is ‘lls’ not ‘ls’. ‘lls’ is for ‘local ls’.
Step # Action
ALERT: It is probably better to check the destination directory using a local file manager or local terminal.
8 Clean up.
sftp> rm *
sftp> cd ..
sftp> rmdir pull
Remote streaming allows operators to stream CSL records to remote server(s) rather than to the
active MGMT VM. This offloads the MGMT VM for high traffic nodes. If remote streaming is not
provisioned, then legacy streaming to the MGMT VM occurs.
Before configuring CSL remote streaming, remote CSL server(s) must first be installed and
commissioned. Once complete, follow these steps to configure CSL remote streaming at the
MME/SGSN node. Note that legacy streaming to the MGMT VM is disabled once the streaming
component is configured.
it should be done during a maintenance window since CSL records may be dropped while
remote-server connections are established
the MGMT VM will only generate CSL records if streaming-fallback is enabled and there
are connection issues with the remote streaming server
if the agwCmnCsl service remains provisioned, then empty CSL files are not generated on the
MGMT VM during periods of inactivity as long as the streaming component remains
provisioned
Step # Action
1 Add the default CSL SCTP profile and if desired, modify the default values.
2 Add the streaming csl component and configure local streaming information. A loopback interface is
Step # Action
required for the CSL IP address.
Syntax:
user@host% set mme-sgsn csl streaming <sctp-profile> <routing-instance> <local-ip-address>
<remote-ip-address1> <remote-port1> <cr or additional parameters>
For example, to provision the local endpoint and one remote server:
sctp-profile csldefaultsctp;
routing-instance EPCCore01;
local-ip-address 172.18.249.154;
remote-ip-address1 172.16.236.179;
remote-port1 30000;
By default, the MME/SGSN does not perform CSL overload record throttling when streaming CSL
records to the active MGMT VM. Follow these steps to limit the maximum number of records that are
written within 10 second intervals. Operational counter ‘records-dropped’ displays the total number of
records dropped due to throttling. This counter is reset to zero anytime the provisioned value ‘max-
records-per-period’ is changed.
Step # Action
1 Modify the default record throttling parameter.
As a fallback mechanism, there is a configurable parameter that allows operators to redirect CSL
records to the MGMT VM when the connection to the remote server is down. By default, if remote
streaming is provisioned and the CSL-Clients are unable to connect to at least one remote CSL
server, then CSL records are discarded until at least one remote connection is established. This
behavior can be overridden to allow fallback to legacy MGMT VM streaming by performing the steps
outlined in the following table. Empty CSL files are not generated on the MGMT VM during periods of
inactivity as long as the streaming component is provisioned.
Note that the agwCmnCsl service must be provisioned before setting ‘streaming-fallback’ to ‘enable’.
Step # Action
1 Modify the streaming fallback parameter.
Two or more (up to 8) LBCtrls can be instantiated on separate hosts. All policies are programmed on
each LB VM allowing the ability to load balance ingress 3GPP traffic across all LB VMs. Egress traffic
is handled on the target VM by routing the packets back to the LB VM using multi-path routing (in
order to load balance the egress 3GPP traffic). 2G and 3G user data does not transverse the LB VM
and is routed directly to the target Data VMs with a static route configured in the next-hop router/L3-
capable switch.
Attention:
It is recommended to have the LB VM interface and the 3GPP interfaces IP addresses in
different subnets.
If 3GPP interfaces share the same subnet as the LB VM, the next hop router sends an ARP
request for the 3GPP interfaces to the LB VM interface but the LB will not respond to the
ARP request.
To allow for the 3GPP interfaces and the LB VM to share the same subnet one of the
following actions can be taken on the next-hop router:
(1) - If the next-hop router supports static routes, then adding static routes for each 3GPP
interface by specifying the destination IP address and the next-hop router IP address
where:
The integrity of the SRM policies is maintained by keeping steering policies in an internal distributed
database to ensure data consistency.
Step # Action
1 Set the steering mechanism to ‘load-balancer’:
Step # Action
user@host% set routing-instances instance EPCAccess01 lb-instance 1 if2-ipv6-addr
fd02::101:0:0:ac1b:6503/64
user@host% set routing-instances instance EPCAccess01 data-instance 0 if1-ipv4-addr
172.27.100.4/24
user@host% set routing-instances instance EPCAccess01 data-instance 0 if1-ipv6-addr
fd02::100:0:0:ac1b:6404/64
user@host% set routing-instances instance EPCAccess01 data-instance 0 if2-ipv4-addr
172.27.101.4/24
user@host% set routing-instances instance EPCAccess01 data-instance 0 if2-ipv6-addr
fd02::101:0:0:ac1b:6504/64
user@host% set routing-instances instance EPCAccess01 data-instance 1 if1-ipv4-addr
172.27.100.5/24
user@host% set routing-instances instance EPCAccess01 data-instance 1 if1-ipv6-addr
fd02::100:0:0:ac1b:6405/64
user@host% set routing-instances instance EPCAccess01 data-instance 1 if2-ipv4-addr
172.27.101.5/24
user@host% set routing-instances instance EPCAccess01 data-instance 1 if2-ipv6-addr
fd02::101:0:0:ac1b:6505/64
user@host% set routing-instances instance EPCCore01 vlan-tag1 200
user@host% set routing-instances instance EPCCore01 vlan-tag2 201
user@host% set routing-instances instance EPCCore01 next-hop-ipv4-gw-addr1 172.27.200.1/24
user@host% set routing-instances instance EPCCore01 next-hop-ipv6-gw-addr1
fd02::200:0:0:ac1b:c801/64
user@host% set routing-instances instance EPCCore01 next-hop-ipv4-gw-addr2 172.27.201.1/24
user@host% set routing-instances instance EPCCore01 next-hop-ipv6-gw-addr2
fd02::201:0:0:ac1b:c901/64
user@host% set routing-instances instance EPCCore01 vlan-host-interface1 eth3
user@host% set routing-instances instance EPCCore01 vlan-host-interface2 eth4
user@host% set routing-instances instance EPCCore01 lb-instance 0 if1-ipv4-addr 172.27.200.2/24
user@host% set routing-instances instance EPCCore01 lb-instance 0 if1-ipv6-addr
fd02::200:0:0:ac1b:c802/64
user@host% set routing-instances instance EPCCore01 lb-instance 0 if2-ipv4-addr 172.27.201.2/24
user@host% set routing-instances instance EPCCore01 lb-instance 0 if2-ipv6-addr
fd02::201:0:0:ac1b:c902/64
user@host% set routing-instances instance EPCCore01 lb-instance 1 if1-ipv4-addr 172.27.200.3/24
user@host% set routing-instances instance EPCCore01 lb-instance 1 if1-ipv6-addr
fd02::200:0:0:ac1b:c803/64
user@host% set routing-instances instance EPCCore01 lb-instance 1 if2-ipv4-addr 172.27.201.3/24
user@host% set routing-instances instance EPCCore01 lb-instance 1 if2-ipv6-addr
fd02::201:0:0:ac1b:c903/64
user@host% set routing-instances instance EPCCore01 data-instance 0 if1-ipv4-addr 172.27.200.4/24
user@host% set routing-instances instance EPCCore01 data-instance 0 if1-ipv6-addr
fd02::200:0:0:ac1b:c804/64
user@host% set routing-instances instance EPCCore01 data-instance 0 if2-ipv4-addr 172.27.201.4/24
user@host% set routing-instances instance EPCCore01 data-instance 0 if2-ipv6-addr
fd02::201:0:0:ac1b:c904/64
user@host% set routing-instances instance EPCCore01 data-instance 1 if1-ipv4-addr 172.27.200.5/24
user@host% set routing-instances instance EPCCore01 data-instance 1 if1-ipv6-addr
fd02::200:0:0:ac1b:c805/64
user@host% set routing-instances instance EPCCore01 data-instance 1 if2-ipv4-addr 172.27.201.5/24
user@host% set routing-instances instance EPCCore01 data-instance 1 if2-ipv6-addr
fd02::201:0:0:ac1b:c905/64
To activate this feature, set the vlan-tagging to ‘disable’, sdn to ‘contrail’ and steering to ‘load-
balancer’. Follow these steps.
Step # Action
1 Turn the feature on feature by setting sdn to contrail, steering to load-balancer and vlan-tagging to
disable.
The following table is provided for guidance regarding valid combinations and usage of the
associated attributes.
The following table provides the mapping between the virtual machine type and the MME applications
that reside on that virtual machine type. This table is all inclusive, meaning it contains both mandatory
and optional elements.
Virtual machines are pre-defined to be home for a set of applications, with some being optional.
Optional processes are those that can be enabled through configuration of a specific interface or the
enabling of a feature through CLI.
The following steps describe the definition of the virtual machines required for a typical MME
configuration.
Step # Action
1 Define the MGMT VM which governs the following processes: CM, PM, FM, CSL, Trace
2 Define the CALLP VM which governs the following processes: SC, SGs, DC, SLs
Step # Action
3 Define the SIG VM which governs the following processes: UPM, S1
user@host% set mme-sgsn system sig unit 0 vm-instance vse01-vm03
4 Define the RM VM which governs the following processes: RM, LI, SBc, UPSM, S1M
user@host% set mme-sgsn system rm unit 0 vm-instance vse05-vm02
user@host% set mme-sgsn system rm unit 1 vm-instance vse06-vm02
The NAS (Non-Access Stratum) EMM (Mobility Management) table is instantiated automatically and
pre-populated with recommended defaults at install time. Most of the default NAS MM parameters do
not need require any adjustment. However, depending on other network criteria, some of these MME
parameters may need adjustment.
Step # Action
1 Optional: Display the default EMM NAS parameters and modify as desired.
Step # Action
imeisv-req-sec-mode-cmd on;
tai-list-max 1;
nas-token-window 5;
ue-service-reject-t3442-value 60;
ta-list-allocation unrestricted;
s1-release-delay-timer disabled;
guti-allocation-intra-tau enable;
voice-paging-t3413 disable;
voice-paging-retry-n3413 0;
eps-services-not-allowed-7
}
[snip]
The NAS (Non-Access Stratum) ESM (Session Management) table is instantiated automatically and
pre-populated with recommended defaults at install time. Most of the default NAS ESM parameters
do not need require any adjustment. However, depending on other network criteria, some of these
MME parameters may need adjustment. In particular, the APN defaults are meant to match what is
configured in the operator’s DNS servers if those features are desired.
Step # Action
1 Optional: Display the default ESM configuration.
esm {
activate-bearer-t3485 5;
activate-bearer-retry-n3485 2;
modify-bearer-t3486 5;
modify-bearer-retry-n3486 2;
esm-info-t3489 4;
esm-info-retry-n3489 3;
deactivate-bearer-t3495 5;
deactivate-bearer-retry-n3495 2;
max-bearers-per-ue 7;
max-pdn-conn-per-ue 8;
deactivate-bearer-page-fail disable;
pgw-ip-type ipv4;
deactivate-bearer-enodeb-fail disable;
}
2 Optional: Modify the ESM parameters.
Step # Action
deactivate-bearer-t3495 5;
deactivate-bearer-retry-n3495 2;
max-bearers-per-ue 7;
max-pdn-conn-per-ue 8;
deactivate-bearer-page-fail disable;
pgw-ip-type ipv4;
deactivate-bearer-enodeb-fail disable;
pcscf-restore-mod-fail-action ignore;
[ok][2015-01-26 07:06:37]
4.1.3 S1 Configuration
Since LTE supports dynamic ENB registration, the minimum configuration required for the S1 is to
create the S1 interface and link the SCTP profile. You can also define the maximum eNBs the system
will support.
Step # Action
1 Configure the SCTP layer.
user@host% set mme-sgsn interface sctp s1defaultsctp rto-initial 500 rto-min 100 rto-max 1000
path-max-retrans 2 association-init-max-transmit 2 association-max-retrans 2 no-delay true max-
out-stream 10 max-in-stream 10 hb-interval 30
2 Define S1 Interface.
Step # Action
wait-for-tau-timer 5;
dscp-mark de;
inter-rat-indirect-forwarding never;
indirect-data-fwd-ip-address 0.0.0.0;
version 10.6.0;
enb-alarm-clear-time 24;
location-reporting-control disable;
fgw-sctp-profile s1defaultsctp;
.... sgw-gtpu-dual-tla-ip-filter enable;
}
[ok][2015-02-17 09:32:59]
In this release the GTP Timers and IP Configurability feature alters GTP functionality such that this
section was created to provide the user with a greater understanding from a complete feature
perspective.
An MME/SGSN tunnels GTP control plane signaling messages to peer nodes over the following
defined interfaces:
S4/S11: Tunnel message to SGW (SGW may use S5/S8 interface to message the PGW, based
on the PLMN)
Gn/Gp: (GnTunnel):
→ Gn for tunnel procedures: gtpV1 tunnel messages to GGSN (GGSN communication may
be on Gn or Gp interface based on the PLMN). (Referred to as GnTunnel)
8.1 feature GTP Timers IP Config S3-S10-S16-Sv) enables the operator to configure and use
separate local IP addresses for the each of the GTP mobility interfaces, namely S3, S10, S16, Gn for
mobility procedures, and for the Sv interface. This functionality was provided for the GTP tunnel
interfaces, namely S4, S11, Gn for tunnel procedures by feature GTP Timers and IP Configurability
feature in 8.0. It allowed a single separate IP for all GTP mobility interfaces as a whole. The feature
provided flexibility to the operator to set different path management timer values like the N3-REQ, T3-
RESP, heartbeat and inactivity timers on a per interface basis for these interfaces. The S101
interface still uses the timers and IP configured in the s101 GTP profile (mme-sgsn interface gtp
profile s101).
The GTP profile is the OAM configuration entity that defines the operational behavior of GTP
interfaces. The profile contains configurations that define IP connectivity information, as well as timing
and reliability parameters that govern the GTP message timers and path management behavior. The
8.0 feature added the capability to designate a GTP profile of these categories:
The 8.0 feature allowed the preceding profile categories to be used for the GTP tunnel interfaces (S4,
S11, Gn Tunnel), and the “default” profile applied to the remaining GTP interfaces (except S101). The
8.1 feature extends this to the GTP mobility interfaces (S3, S10, S16, Gn for mobility messages, and
Sv).
In previous releases, there were three GTP profiles, namely, “tunnel”, “mobility” and “s101”, which
defined the IP connectivity and messaging timers for all GTP based interfaces. The “tunnel” profile
managed the S11, S4 and gnTunnel interfaces. The “mobility” profile managed S10, S3, S16, Sv and
gnMobility interfaces. The “s101” profile managed the S101 interface.
The “tunnel” and “mobility” profiles were mandatory configurations required for the system to function.
The 8.0 feature replaced the above mentioned three profiles with a mandatory “default” profile. The
“s101’ profile remains unchanged. All other gtp-profiles are referred to as generic gtp-profiles, which
may be mapped as interface specific or PLMN specific profiles.
Operators have the option of configuring profiles on a per interface basis to customize the timers and
IP addresses for each of the tunnel interfaces. The following config entities were defined in 8.0 and
used:
s11-if
s4-if
gn-if
The 8.1 feature enhances this by providing the option of configuring profiles on a per interface basis
for the mobility interfaces and the Sv interface. In order to do this, the following config entities are
defined and used:
s3-if
s10-if
s16-if
sv-if
gn-if
Each of the preceding config entities have “gtp-profile” attributes, which can be linked to any generic
gtp-profile. If a generic profile has been linked to one of these interface entities, then any messaging
on that interface will use the IP and routing parameters and associated messaging and path
management timer values as configured in the linked generic profile.
Up to 32 GTP profiles may be configured in the system, including the default gtp-profile and the s101
gtp-profile. This leaves operators a choice of up to 30 generic GTP profiles that may be configured for
use as interface-based profiles for tunnel and s101 interfaces or PLMN based timer profiles.
Network operators generally have roaming agreements with several partners, enabling UEs to roam
across multiple PLMNs. This can lead to network latency issues in certain cases based on the
topological location of the nodal elements across the different PLMNs. The 8.0 feature helped to
alleviate this by allowing operators to use a different set of N3-REQ/T3-RESP timer values for GTP
tunnel interface messaging across different partner PLMNs for the GTP tunnel interfaces. The 8.1
feature extends this to the GTP mobility interfaces. S101 interface is not affected by this
configuration.
This is achieved by linking a configured generic gtp-profile name to a specific plmn-id in the
subscriber-plmn table. When the MME/SGSN detects that GTP messaging being routed across a
PLMN which has a linked gtp-profile, then the N3-REQ/T3-RESP values are selected from the linked
profile. However, the local sender IP/Routing-instance settings are used from the interface based
profiles.
The 8.0 feature relocated globally applicable GTP attributes such as max-pending-transactions,
gtpv1-version and gtpv2-version which were present in the gtp-profile to a parent gtp entity. These
global attributes apply to messaging on all GTP based interfaces.
Step # Action
1 Configure GTP common profile parameters.
Step # Action
user@host% set mme-sgsn interface gtp profile default ip-address-v4 192.168.40.233 routing-
instance EPCCore01 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-
request 8 t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-
response-session-creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5 n3-
request-context-request 2 t3-Response-context-request 5 n3-request-forward-relocation 2 t3-
response-forward-relocation 5
user@host% set mme-sgsn interface gtp profile profile1 ip-address-v4 192.168.40.234 routing-
instance EPCCore01 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-
request 8 t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-
response-session-creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5 n3-
request-context-request 2 t3-Response-context-request 5 n3-request-forward-relocation 2 t3-
response-forward-relocation 5
user@host% set mme-sgsn interface gtp profile profile2 ip-address-v4 192.168.40.235 routing-
instance EPCCore01 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-
request 8 t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-
response-session-creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5 n3-
request-context-request 2 t3-Response-context-request 5 n3-request-forward-relocation 2 t3-
response-forward-relocation 5
To configure the S10/S3 and S11 interface’s you will create two GTP profiles. A tunnel profile for S11,
and a mobility profile for S10/S3.
Step # Action
1 Configure the GTP common profile parameters.
Step # Action
user@host% set mme-sgsn interface gtp profile default routing-instance EPCCore01 ip-address-v4
172.16.44.3 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-request 8
t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-response-session-
creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5 n3-request-context-
request 2 t3-response-context-request 5 n3-request-forward-relocation 2 t3-response-forward-
relocation 5
Optionally, if separation of S4 and S11 source IP addresses is desired, configure a profile specific to the S11
interface or S4 interface and link it under that interface. This will override the implied reference to the generic tunnel
profile for that particular interface.
user@host% set mme-sgsn interface gtp profile s11 ip-address-v4 1.2.3.4 routing-instance
EPCCore01
user@host% set mme-sgsn interface s11 s11-if gtp-profile s11
user@host% set mme-sgsn interface gtp profile mobility routing-instance EPCCore01 ip-address-v4
172.16.44.3 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-request 8
t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-response-session-
creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5 n3-request-context-
request 2 t3-response-context-request 5 n3-request-forward-relocation 2 t3-response-forward-
relocation 5
Unless specified, the Tunnel and Mobility IPs can be the same.
Step # Action
n3-request-bearer-modification 2;
t3-request-bearer-modification 5;
n3-request-context-request 2;
t3-response-context-request 5;
n3-request-forward-relocation 2;
t3-response-forward-relocation 5;
Step # Action
1 Configure GTP Default profile.
user@host% set mme-sgsn interface gtp profile default ip-address-v4 192.168.40.233 routing-
instance EPCCore01 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-
request 8 t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-
response-session-creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5
n3-request-context-request 2 t3-Response-context-request 5 n3-request-forward-relocation 2 t3-
response-forward-relocation 5
user@host% set mme-sgsn interface gtp profile profile1 ip-address-v4 192.168.40.234 routing-
instance EPCCore01 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-
request 8 t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-
response-session-creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5
n3-request-context-request 2 t3-Response-context-request 5 n3-request-forward-relocation 2 t3-
response-forward-relocation 5
user@host% set mme-sgsn interface gtp profile profile2 ip-address-v4 192.168.40.235 routing-
instance EPCCore01 ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-
request 8 t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-
response-session-creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5
n3-request-context-request 2 t3-Response-context-request 5 n3-request-forward-relocation 2 t3-
response-forward-relocation 5
user@host% set mme-sgsn interface gtp profile profile3 ip-address-v4 0.0.0.0 routing-instance
NULL ip-address-v6 :: dscp-mark de echo-interval 300 t3-echo-response 15 n3-echo-request 8 t3-
response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-response-session-
Step # Action
creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5 n3-request-
context-request 2 t3-Response-context-request 5 n3-request-forward-relocation 2 t3-response-
forward-relocation 5
NOTE: Unlock the rm VM pair, then the callp VM pairs, then the sig VMs, and lastly the data VMs.
user@host% request mme-sgsn system rm unlock unit-id <a>
user@host% request mme-sgsn system rm unlock unit-id <b>
user@host% request mme-sgsn system callp <x> unlock unit-id <y>
user@host% request mme-sgsn system callp <x> unlock unit-id <z>
user@host% request mme-sgsn system sig unlock unit-id <m>
user@host% request mme-sgsn system data unlock unit-id <n>
Step # Action
1 Define the MME Pools.
user@host% set mme-sgsn service-area mme-pool pool2 mme-name mme78 mme-groupid 32768 mme-code 64
relative-capacity 100
mme-name mme78;
mme-groupid 32768;
mme-code 64;
relative-capacity 100;
ALERT: To avoid STMSI collisions within a node, configure unique mme-codes across all mme-pools.
user@host% set mme-sgsn service-area plmn 10068 mme-pool pool1 sgsn-pool default
Step # Action
user@host% show mme-sgsn service-area plmn 10068
mme-pool pool1;
sgsn-pool default;
Step # Action
1 Configure the SCTP layer.
user@host% set mme-sgsn interface sctp s6defaultsctp rto-initial 500 rto-min 100 rto-max 1000
path-max-retrans 2 association-init-max-transmit 2 association-max-retrans 2 no-delay true max-
out-stream 10 max-in-stream 10 hb-interval 30
Step # Action
association-max-retrans 2;
no-delay true;
max-out-stream 10;
max-in-stream 10;
hb-interval 30;
user@host% set mme-sgsn interface diameter profile sctp-profile s6defaultsctp mme-realm affirmed-
cta.com watchdog-timer 30 reconnect-timer 30 peer-state-timer 30 dscp-mark de peer-selection-alg
load-based
4 Define initial diameter client (local). A loopback interface is required for the diameter client IP address.
user@host% set mme-sgsn interface diameter peer <host-identity> ip-address 172.16.42.2 ip-
address2 0.0.0.0 port-number 3868
Step # Action
6 Map the realm to the Diameter remote peer.
8 Modify the HLR/HSS route to enable routing options of: s6d-override, gr-fallback, s6d-override-and-gr-
fallback, standard. The default is standard.
user@host% set mme-sgsn imsi-routing hss-hlr-routing 45005 routing-option s6d-override
OR
user@host% set mme-sgsn imsi-routing hss-hlr-routing 45005 routing-option gr-fallback
OR
user@host% set mme-sgsn imsi-routing hss-hlr-routing 45005 routing-option s6d-override-and-gr-
fallback
OR
user@host% set mme-sgsn imsi-routing hss-hlr-routing 45005 routing-option ue-epc-capability
To remove any routing option provisioned, set the routing option to standard:
user@host% set mme-sgsn imsi-routing hss-hlr-routing 45005 routing-option standard
The MME predicts the tracking area a UE is likely going to travel into, based on the mobility
information of all UEs on the MME.
By including the likely TAs in the TAI list for a UE, the UE does not have to perform a tracking
area update when it moves into a tracking area in the list, thereby minimizing the non-revenue
generating control signaling.
A UE that has a high probability of being paged relative to its tracking area updates will have a
smaller tracking area list, since the cost of paging is typically higher than a single tracking area
update.
Step # Action
When enabled, MME will use the TA probability computation to generate a UEs TA list.
When disabled, MME will create the TA probability table, but not use it for a UEs TA list generation.
NEIGHBORTAI
PLMN PERCENT
ID TAC SGWRELOC
----------------------------
50502 0x1001 -
50502 0x1002 -
50502 0x1003 -
50502 0x1004 -
50502 0x1006 -
50502 0x1118 -
50502 0x1145 -
50502 0x2001 -
50502 0x2002 -
50502 0x2118 -
50502 0x31a4 -
50502 0x41a4 -
50502 0x5152 -
50502 0x5154 -
50502 0x8001 -
50502 0x8002 -
50502 0x8003 -
50502 0x8004 -
50502 0x9001 -
50502 0x9002 -
50631 0x1001 -
50631 0x1002 -
50631 0x1003 -
50631 0x1004 -
50631 0x1006 -
Step # Action
50631 0x1118 -
50631 0x1145 -
50631 0x2001 -
50631 0x2002 -
50631 0x2118 -
50631 0x31a4 -
50631 0x41a4 -
50631 0x5152 -
50631 0x5154 -
50631 0x8001 -
50631 0x8002 -
50631 0x8003 -
50631 0x8004 -
50631 0x9001 -
50631 0x9002 -
[ok][2014-12-11 11:20:32]
4.2.2 Trace
The MME supports trace of external messaging on a per UE basis with output in XML file format as
defined by 3GPP standards 32.421, 32,422, 32.423. Basic configuration is shown in the following.
Please refer to the Monitoring Guide for more detailed examples of using this feature. 3GPP Trace
functionality runs on two different VMs, the CALLP VM, where the trace data is collected and the
MGMT VM, where the trace files are written.
Step # Action
1 Enable the feature.
Step # Action
user@host% set mme-sgsn trace session <target-type> imei 210000011000060 12345.123411 s1 att-tau-
det s1 s11 bearer-act-mod-del off none med 1.2.3.4
user@host% set mme-sgsn trace session <target-type> imei 210000011000060 12345.123411 s1,s10,s11
att-tau-det s1, x2,uu s11 bearer-act-mod-del off none med 1.2.3.4
user@host% set mme-sgsn trace session <target-type> imei 210000011000060 12345.123411 s1,s10,s11
att-tau-det s1 s11 bearer-act-mod-del off none min 1.2.3.4
For deleting:
user@host% delete mme-sgsn trace session <target-type> imsi 312980911000003
user@host% delete mme-sgsn trace session <target-type> imeisv 1234567892222202
user@host% delete mme-sgsn trace session <target-type> imei 210000011000060
user@host> shell
user@host:~$ cd /usr/local/epc/trace/
user@host:/usr/local/epc/trace$ ls -ltr open closed
The File is 3GPP defined XML format and may be transferred off-board by SFTP pull.
Step # Action
1 Add the S11 interface table to control any optional features for that interface, or to point to a unique gtp-
profile for this particular interface when required
The change-reporting parameter indicates whether MME/SGSN provides user location information (ULI) for every
mobility event.
The uli-inclusion parameter controls if MME will send ULI in the below GTPv2 messages if it is set to 'force-in-all'
• Create Session Request(HO/TAU)
• Modify Bearer Request
• Delete Session Request
• Delete Bearer Command
And if it is set to 'force-in-mbr', then ULI will be included in every Modify Bearer Request message when requested by
GW.
Configurable parameter to control the use of PDN Gateway Identity (PGWid) received from HSS
Step # Action
1 Enable the PGW selection feature and adjust reselection retry count (default off, and 2 respectively).
Step # Action
Step # Action
1 Enable the feature under the s11 interface config.
2 Unless the functionality is also enabled on SGW and PGW, no behavior differences will occur.
Check feature support and configuration on SGW and PGW nodes.
Step # Action
1 Ensure adjacent nodes are configured correctly for the desired PGW.
• DNS: The operator must provision the PDN GW on the DNS server for use by MVNO PDN with only the x-s5-gtp
app-protocol.
Step # Action
• HSS: The HSS must not send an IP address for the PDN GW Identity for an APN that may be used for a MVNO
PDN.
• FXA: The AAA (Fxa Interface) must not send an IP address for the PDN GW Identity for an APN that may be used
for a MVNO PDN.
2 Provision a bearer-limits-profile for MVNO default bearers (ALERT: Provision QCIs in the range of 5-9).
Syntax:
user@host% set mme-sgsn subscriber bearer-limits-profile mvno-default qos-class-identifier <qos-
class-identifier> arp-priority <arp-priority>
3 Provision a bearer-limits-profile for MVNO dedicated bearers (Note: Provision QCIs in the range of 1-9):
Syntax:
user@host% set mme-sgsn subscriber bearer-limits-profile mvno-dedicated qos-class-identifier
<qos-class-identifier> arp-priority <arp-priority> max-bitrate-uplink <max-bitrate-uplink> max-
bitrate-downlink <max-bitrate-downlink> guaranteed-bitrate-uplink <guaranteed-bitrate-uplink>
guaranteed-bitrate-downlink <guaranteed-bitrate-downlink>
To configure the S101 interface you will first create the optional S101 GTP profile.
For a greater understanding of GTP profile usage, see GTP Feature Overview.
Step # Action
1 Verify the GTP common profile parameters.
user@host% set mme-sgsn interface gtp profile s101 routing-instance EPCCore01 ip-address-v4
172.16.44.3 ip-address-v6 :: dscp-mark de echo-interval 4 t3-echo-response 15 n3-echo-request 8
t3-response 5 n3-request 2 inactivity-timer 10 n3-request-session-creation 2 t3-response-session-
creation 5 n3-request-bearer-modification 2 t3-request-bearer-modification 5 n3-request-context-
request 2 t3-response-context-request 5 n3-request-forward-relocation 2 t3-response-forward-
relocation 5
s101 {
routing-instance EPCCore01;
ip-address-v4 172.16.44.3;
ip-address-v6 ::;
dscp-mark de;
echo-interval 4;
t3-echo-response 15;
n3-echo-request 8;
t3-response 5;
n3-request 2;
inactivity-timer 10;
n3-request-session-creation 2;
t3-response-session-creation 5;
n3-request-bearer-modification 2;
t3-request-bearer-modification 5;
n3-request-context-request 2;
t3-response-context-request 5;
n3-request-forward-relocation 2;
t3-response-forward-relocation 5;
}
Step # Action
}
[snip]
Step # Action
1 Enable the feature and configure the interface.
user@host% set mme-sgsn interface s102 iws 0x0102 ip-address 172.16.241.148 port 4598
3 Show the eNB to IWS mapping (discovered, persistent, but not configured on MME).
Supports up to 15000 entries, so selective command variants are helpful.
Step # Action
1 Ensure the S1 interface is configured for 9.5.1 version of the standards.
routing-instance1 EPCAccess01;
routing-instance2 EPCAccess02;
sctp-profile s1defaultsctp;
ip-address1-v4 172.18.249.1;
ip-address2-v4 172.18.249.11;
ip-address1-v6 ::;
ip-address2-v6 ::;
port 36412;
time-to-wait 60;
reloc-prep-timer 3;
reloc-overall-timer 5;
indirect-tunnel-hold-timer 5;
handover-resource-release-timer 5;
ue-context-release-timer 5;
ue-context-modification-timer 5;
wait-for-tau-timer 5;
dscp-mark de;
inter-rat-indirect-forwarding never;
indirect-data-fwd-ip-address 0.0.0.0;
version 9.5.1;
enb-alarm-clear-time 24;
location-reporting-control disable;
lrc-timer-direct 5;
csg-expiration-release-timer 5;
sgw-gtpu-dual-tla-ip-filter enable
routing-instance EPCCore01;
hss-realm-for-min-query affirmed-cta.com;
ip-address-v4 172.18.249.81;
ip-address-v6 ::;
port 23272;
version icd-v2;
dscp-mark de;
echo-interval 4;
t-ack-21-timer 5;
request-retry 1;
max-pending-transactions 250;
priority-paging enable;
enb-map-lifetime 24;
Step # Action
Step # Action
1 Configure the interface.
Syntax:
user@host% set mme-sgsn interface s13 s13-if realm-name <realm-name> request-timer <request-
timer> request-retry <request-retry> include-imsi <include-imsi> eir-host-id <eir-host-id>
version <version>
user@host% set mme-sgsn interface diameter peer foo-eir ip-address 1.2.3.4 ip-address2 0.0.0.0
port-number 3868
Step # Action
1 Add an AVP for which M-bit override is desired for s6, s13, or base application.
Step # Action
1 Configure the SCTP profile.
user@host% set mme-sgsn interface sctp sctpSgs rto-initial 500 rto-min 250 rto-max 2000 path-max-
retrans 2 association-init-max-transmit 2 association-max-retrans 2 no-delay true max-out-stream
1 max-in-stream 1 hb-interval 30 rx-buffer-size 149000 tx-buffer-size 149000 sack-delay 200 sack-
frequency 2
2 Enable the feature and configure the SGs interface. A loopback interface is required for the SGs
interface IP address.
Step # Action
routing-instance1 EPCCore01;
routing-instance2 EPCCore02;
ip-address 172.18.249.27;
sctp-profile sctpSgs;
loc-upd-guard-ts6-1 10;
eps-detach-guard-ts8 4;
explicit-imsi-detach-guard-ts9 4;
implicit-imsi-detach-guard-ts10 4;
mme-reset-ind-guard-ts12-2 4;
loc-upd-retry-ns6 2;
eps-detach-retry-ns8 2;
explicit-imsi-detach-retry-ns9 2;
implicit-imsi-detach-retry-ns10 2;
mme-reset-ind-retry-ns12 2;
sms-only false;
mme-reset-paging-type cs-paging;
mme-reset-indication true;
version 9.3.0;
dscp-mark de;
ip-address2 0.0.0.0;
3 Configure up to 64 VLRs.
ALERT: The length of the name should be limited to 31 chars so that it matches the name configured
under vlr-pool in the next step. Although the standards recommend an FQDN for the name, it actually is
not sent in any message for MME/SGSN and only has internal significance to link these two tables
together.
Note: Attribute vlr-type can be: Gs, SGs, Sv or combination of interfaces. For a VLR that supports more
than one type, input one set command for each type, but use the same VLR name each time.
user@host% set mme-sgsn service-area vlr sgsserver.affirmed.com vlr-type sgs vlr-number 1 callp-
service-id 2 sgs-ip-address1 172.16.236.11 sgs-ip-address2 0.0.0.0
4 Configure the default VLR for the VLR pool, up to 32 VLR pools are supported.
5 (Optional) Configure the VLR pool with the selected VLR for an IMSI hash range, up to 32 non-
overlapping IMSI hash ranges are supported for a VLR pool.
Note: The MSC selection is not based on the actual IMSI digits, but on the IMSI hash result specified in TS 23.236.
The MME first does the IMSI hash and then uses that output to select a VLR from the VLR pool: ((IMSI div 10)
modulo 1000) Which reveals a number between the range 0 to 999.
The configuration vlr-selection-digits is the actual range of IMSI hash that is mapped to a particular vlr-pool
0..200 => POOL1 201..599 => POOL2 600..999 => POOL3
This ultimately translates to the use of the digits of the IMSI: 123451234512xxx5
Step # Action
6 Configure the Location Area.
Syntax:
set mme-sgsn service-area location-area <plmn-id> <lac> vlr-pool-name <vlr pool name>
4.2.13 Sv to VLR
The Sv interface supports the Single Radio Voice Call Continuity procedure between the MME and
the VLR such that a voice call started on the EPC network using VoLTE can be handed over to the
legacy circuit domain seamlessly.
Follow these steps to configure the Sv interface IP address of the VLR and to set the type of VLR
discovery (either local or DNS).
Step # Action
1 Configure the VLR discovery type (local or DNS).
Note: If the attribute value is ‘dns’ then the MME uses DNS query to look up the MSC server.
Note: If the attribute value is ‘local-config’ then the MME uses the locally configured Sv interface ip-
address and bypasses the DNS query.
Step # Action
1 Create a sctp profile, and modify the default buffer sizes to be larger than default values.
user@host% set mme-sgsn interface sctp sctpSbc rto-initial 500 rto-min 250 rto-max 2000 path-max-
retrans 2 association-init-max-transmit association-max-retrans 2 no-delay true max-out-stream 1
max-in-stream 1 hb-interval 30 rx-buffer-size 200000 tx-buffer-size 200000 sack-delay 200 sack-
frequency 2
[ok][ 2014-06-22 23:10:02]
2 Enable the feature and create the SBc interface. A loopback interface is required for the SBc interface
IP address. In this example, the sbc-type is set to ‘pws’ (vs the other option of ‘etws’).
Step # Action
3 Display warning message retries and retry timer.
Step # Action
1 Activate the FXA feature.
2 Define up to 4096 APN NI that will trigger the FXA based P-GW selection procedure.
user@host% set mme-sgsn interface fxa fxa-if routing-instance EPCCore01 ip-address 172.18.249.88
ports 36664,36665 path-mgmt-user-name qwerty path-mgmt-user-password qwertyqwertyqwerty path-
mgmt-imsi 1234567890123456
Step # Action
user@host% set mme-sgsn interface radius peer 172.16.241.131 port-number 1812 application-
identifier fxa
If the guti-lifetime has expired for the UE, the MME will allocate a new GUTI during intra-MME TAU
regardless of the setting of the new option.
Step # Action
1 Disable GUTI allocation during TAU (enabled by default), a NAS EMM function.
ue-retry-delay-t3402-value 12;
ue-periodic-tau-t3412-value 54;
paging-t3413-initial 5;
paging-t3413-increment 2;
paging-retry-n3413 2;
targeted-first-page all-ta;
nwk-init-detach-t3422 6;
nwk-init-detach-retry-n3422 2;
guti-allocation-t3450 6;
guti-allocation-retry-n3450 3;
guti-lifetime 12;
security-t3460 6;
Step # Action
security-retry-n3460 2;
identity-t3470 6;
identity-retry-n3470 2;
mobile-reachable-timer t3412plus4;
opt-authentication-control off;
vector-retrieval-failure ignore;
ciphering-algorithms eea0,eea1,eea2;
integrity-algorithms eia1,eia2,null;
src-mme-resource-release-timer 15;
cancel-location-wait-timer 10;
imeisv-req-sec-mode-cmd on;
tai-list-max 1;
nas-token-window 5;
ue-service-reject-t3442-value 60;
ta-list-allocation unrestricted;
s1-release-delay-timer disabled;
guti-allocation-intra-tau disable;
voice-paging-t3413 disable;
voice-paging-retry-n3413 0;
Step # Action
1 Setting the default EMM cause code and error description mapping. Also displaying the cause code
value.
user@host% set mme-sgsn interface nas emm-cause-mapping user-unknown cause-code eps-service-not-
allowed-in-this-plmn-1 unknown-scptn-no-gprs-data cause-code eps-service-not-allowed-7 unknown-
scptn-gprs-data cause-code no-suitable-cells-in-ta-15 rat-not-allowed cause-code eps-service-not-
allowed-in-this-plmn-14 roaming-not-allowed cause-code eps-service-not-allowed-in-this-plmn-14
equipment-unknown cause-code eps-service-not-allowed-in-this-plmn-1 authentication-data-
unavailable cause-code network-failure-17 access-restricted cause-code eps-service-not-allowed-
in-this-plmn-14 operator-determined-barring cause-code eps-service-not-allowed-in-this-plmn-14
sm-layer-failure cause-code esm-failure-19 unknown-hss cause-code eps-service-not-allowed-in-
this-plmn-14 dns-mapping-failure cause-code ue-id-cannot-be-derived-9 ta-restricted cause-code
roaming-not-allowed-in-this-ta-13 authorization-rejected cause-code no-suitable-cells-in-ta-15
application-unsupported cause-code no-suitable-cells-in-ta-15 command-unsupported cause-code no-
suitable-cells-in-ta-15 out-of-space cause-code congestion-22 too-busy cause-code congestion-22
unable to comply cause-code no-suitable-cells-in-ta-15 unknown-serving-node cause-code imsi-
unknown-in-hss-2 roaming-restricted-in-mme-due-to-unsupported-feature cause-code no-suitable-
cells-in-ta-15
user@host% show mme-sgsn interface nas emm-cause-mapping access-restricted
cause-code eps-service-not-allowed-in-this-plmn-14;
In the preceding command, error description should be chosen from the listed string values and cause code can be
any of the listed strings or a numeric value in the range of 1-255.
2 Modify the EMM cause code and error description mapping.
user@host% set mme-sgsn interface nas emm-cause-code-mapping <error> cause-code <cause code>
Example:
user@host% set mme-sgsn interface nas emm-cause-code-mapping access-restricted cause-code
congestion-22
In the preceding command, error description should be chosen from the listed string values and
cause code can be any of the listed strings or a numeric value in the range of 1-255.
Step # Action
1 Setting an Emergency Number into the list.
EMM Information is an optional message that is sent from the MME to the UE so the UE can update
the network name and time information. The EMM Information message is by default only sent by the
MME if initiated by the VLR. This feature adds the ability for the MME to send the EMM Information
message on any combination of the following provisioned triggers:
Attach
Inter-MME TAU
For the provisioned triggers above, the EMM Information message will include the locally provisioned
information. For the VLR initiated trigger, the EMM Information message will include the information
from the VLR. The usage of the information message can be measured by the following PM counter:
lteMm::procedure::VS.MM.EmmInformation
Step # Action
1 Enable EMM Information to any one of the triggers (example: dst-change).
Step # Action
1 Setting the default configuration for paging overload.
user@host% set mme-sgsn engineering paging-overload minor-ovld enable minor-onset 325 minor-abate
300 paging-delay-minor 10 paging-reject-minor 0 major-ovld enable major-onset 375 major-abate 350
paging-delay-major 20 paging-reject-major 0 critical-ovld enable critical-onset 425 critical-
abate 400 paging-delay-critical 40 paging-reject-critical 0
Step # Action
paging-reject-critical 0;
Step # Action
1 Modify and display the SC engineering configuration details.
Step # Action
user@host% set mme-sgsn engineering sc-engineering pdn-average 105 dedicated-bearer-average 200
apn-average 200 csg-average 200
Step # Action
1 Configure the gtp mobility profile version to release 10.
user@host% set mme-sgsn interface gtp gtpv1-version 10.8.0
Step # Action
[ok][2014-05-23 10:00:45]
A UE can request for emergency services using one of two available procedures:
Step # Action
1 Verify that attributes ‘max-pdn-conn-per-ue’ and ‘max-bearers-per-ue’ in mme-sgsn interface nas esm
are at least 2.
Use the modify command if they are not.
user@host% set mme-sgsn interface nas esm max-bearers-per-ue 7 max-pdn-conn-per-ue 8
user@host% show mme-sgsn interface nas esm
activate-bearer-t3485 5;
activate-bearer-retry-n3485 2;
modify-bearer-t3486 5;
modify-bearer-retry-n3486 2;
esm-info-t3489 4;
esm-info-retry-n3489 2;
deactivate-bearer-t3495 5;
deactivate-bearer-retry-n3495 2;
max-bearers-per-ue 7;
max-pdn-conn-per-ue 8;
deactivate-bearer-page-fail disable;
pgw-ip-type ipv4;
deactivate-bearer-enodeb-fail disable;
Step # Action
1 Verify that attributes ‘max-pdn-conn-per-ue ‘and ‘max-bearers-per-ue’ in interface nas esm are at least
2.
Use the modify command if they are not.
user@host% set mme-sgsn interface nas esm max-bearers-per-ue 7 max-pdn-conn-per-ue 8
user@host% show mme-sgsn interface nas esm
activate-bearer-t3485 5;
activate-bearer-retry-n3485 2;
modify-bearer-t3486 5;
modify-bearer-retry-n3486 2;
esm-info-t3489 4;
esm-info-retry-n3489 2;
deactivate-bearer-t3495 5;
deactivate-bearer-retry-n3495 2;
max-bearers-per-ue 7;
max-pdn-conn-per-ue 8;
deactivate-bearer-page-fail disable;
pgw-ip-type ipv4;
deactivate-bearer-enodeb-fail disable;
2 Set the following interface versions to Rel 10: s1-if, s6-if, gtp and sgs-if
When the UE attaches to the network, as part of update location request (ULR) the NPLI capabilities
of the MME/SGSN is provided to the HSS, if the feature is turned on. The MME/SGSN indicates
support of State/Location Information Retrieval and UE time zone retrieval for the UE.
When the MME/SGSN supports NPLI retrieval, the HSS will query for user state, location information,
UE time zone information in a subsequent Insert Subscriber Data Request (IDR) towards that
MME/SGSN. The MME/SGSN will send the response back to HSS with the UE state, location
information, UE Time Zone in Insert Subscriber Data response (IDA).
User State
E-UTRAN-Cell-Global-Identity (eCGI)
Tracking-Area-Identity (TAI)
Cell-Global-Identity (CGI)
Location-Area-Identity (LAI)
Service-Area-Identity (SAI)
Routing-Area-Identity (RAI)
Age of Location
UE Time zone
Step # Action
1 Verify/set the s6-if to Rel 10.
3 (optional) Enable ue-timezone to support an MME with UE Time Zone reporting towards the HSS.
The SLs process is added for the SLs interface. This is an optional process that resides in a CallP VM
together with the subscriber control processes. It is created when:
In the following activation steps, we assume that a callp VM is already established and a diameter
process is already added.
Step # Action
1 Add the slg-if:
Syntax:
user@host% set mme-sgsn interface slg slg-if gmlc-realm <gmlc realm> request-timer <timer value>
request-retry <retry times> version <version>
To use predefined default values:
Examples:
user@host% set mme-sgsn slg slg-if gmlc-realm affirmed-cta.com request-timer 5 request-retry 0
version 10.1.0
user@host% set mme-sgsn interface slg gmlc <gmlc-number> host-identity <host identity>
Examples:
user@host% set mme-sgsn interface slg gmlc 460161003029 host-identity host1 gmlc-usage emergency-
Step # Action
regular
Examples:
user@host% set mme-sgsn interface sctp sctpSls rto-initial 500 rto-min 250 rto-max 2000 path-max-
retrans 2 association-init-max-transmit 2 association-max-retrans 2 no-delay true max-out-stream
1 max-in-stream 1 hb-interval 30 rx-buffer-size 149000 tx-buffer-size 149000 sack-delay 200 sack-
frequency 2
Step # Action
Example:
user@host% set mme-sgsn interface sls sls-if routing-instance1 EPCCore01 routing-instance2 0
sctp-profile sctpSls local-ip-address1 192.168.40.23 local-ip-address2 0.0.0.0 port 9082
location-request-t3x01 60 reset-request-t3x02 3 reset-request-retry 0 version 10.4.0 dscp-mark de
max-transactions-per-peer 5000
Step # Action
user@host% show mme-sgsn interface sls e-smlc
e-smlc 1 {
callp-service-id 0;
e-smlc-pool pool1;
ip-address1 172.16.236.63;
}
8 Modify the subscriber-plmn:
Syntax:
user@host% set mme-sgsn subscriber subscriber-plmn <default|plmn-id> mt-lr <enable|disable> mo-lr
<disable>
To use predefined default values of service-area:
user@host% set mme-sgsn subscriber subscriber-plmn
Examples:
user@host% set mme-sgsn subscriber subscriber-plmn 31298 mt-lr enable mo-lr disable
name lcs;
loc-notification-timer-tlcsn 15;
lcs-over-sgs disable;
If you wish to change one of them, use one of the following commands:
Step # Action
user@host% set mme-sgsn feature lcs loc-notification-timer-tlcsn 20
user@host% commit
Gn-SGSN refers to an SGSN which supports the Gn interface to GGSN, but does not support the
S4 interface to SGW.
S4-SGSN refers to an SGSN which supports the S4 interface, but does not support the Gn
interface.
Simply SGSN refers to an SGSN supporting both the S4 and the Gn interfaces.
An S4-SGSN may optionally use S6d and S13 interfaces as well, or continue with Gr and Gf
interfaces for older networks. The S3 and S16 interfaces are mobility variants of GTP path, and the
S4 interface is a tunnel variant of GTP path. The S4-SGSN requires R9 or greater standard
compliancy on the relevant interfaces; R8 variants are not supported.
The following table provides the mapping between the virtual machine type and the SGSN
applications that reside on that virtual machine type. This table is all inclusive, meaning it contains
both mandatory and optional elements.
DATA SD, GB
SLB LB
Virtual machines are pre-defined to be home for a set of applications, with some being optional.
Optional processes are those that can be enabled through configuration of a specific interface or the
enabling of a feature through CLI.
The following steps describe the definition of the virtual machines required for a typical SGSN
configuration.
Step # Action
1 Define the MGMT VM which governs the following processes: CSL, Trace, CM, PM, FM
2 Define the CALLP VM which governs the following processes: SC, SGS, DC, Iu, Ge
3 Define the SIG VM which governs the following processes: UPM, ASP, IPSP, TCAP
5 Define the RM VM which governs the following processes: RM, LI, SBc, Ga, UPSM
The setting of NAS GMM is shown in the following table. The config values assigned are default
values. They need to be modified for specific network operation.
Step # Action
1 Set the GMM NAS parameters.
5.1.2.2 NAS SM
The setting of NAS SGSN SM is shown in the following table. The config values assigned are default
values and they do not require adjustment. However, depending on other network criteria, some of
these SGSN parameters may need adjustment. In particular, the APN defaults are meant to match
what is configured in the operator’s DNS servers if those features are desired.
Step # Action
1 Set the SM NAS parameters.
Show the defaults after new install:
user@host% show mme-sgsn interface nas sm
nwk-pdp-deact-timer-t3395 8;
nwk-pdp-deact-retry-n3395 4;
ppp-def-apn-network-id default.ppp-apnni.com;
ip-def-apn-network-id default.ip-apnni.com;
def-apn-operator-id default.apnoi.gprs;
nwk-pdp-modify-timer-t3386 8;
nwk-pdp-modify-retry-n3386 4;
pdp-activations-per-mobile 7;
ipv6-enduser-address disable;
route-advance enable;
qos-negotiation disable;
qos-upgrade-support enable;
local-qos-control never;
old-sgsn-buffering disable;
To configure the Gr interface, you will need to link the SCTP/M3UA information to the local SS7
information. The rest of the Gr interface will be configured based on the network design. At a
minimum, one remote peer and GTT transaction needs to be configured to route messages external
and internal to the SGSN.
Step # Action
1 Configure the SCTP layer.
user@host% set mme-sgsn interface sctp sctpSgsn rto-initial 500 rto-min 100 rto-max 1000 path-
max-retrans 2 association-init-max-transmit 2 association-max-retrans 2 no-delay true max-out-
stream 10 max-in-stream 10 hb-interval 30
user@host% show mme-sgsn interface sctp sctpSgsn
rto-initial 500;
rto-min 100;
rto-max 1000;
path-max-retrans 2;
association-init-max-transmit 2;
association-max-retrans 2;
no-delay true;
max-out-stream 10;
max-in-stream 10;
hb-interval 30;
user@host% set mme-sgsn interface m3ua m3uaSigtran sctp-profile sctpSgsn audit-interval 30000
multiple-network-appearances enable max-deactivate-attempts 3 fsm-timeout 2000 network-indicator
natl cong-onset-level1 auto-configure cong-abatement-level1 auto-configure cong-onset-level2
auto-configure cong-abatement-level2 auto-configure cong-onset-level3 auto-configure cong-
abatement-level3 auto-configure cong-report-frequency 8 reconnect-timer 10
3 Configure the local SS7 pointcode and E.164 number. A loopback interface is required for the ASP IP
address.
user@host% set mme-sgsn interface ss7 ss7 sgsn-number 601001001001 signaling-variant ansi map-
cgpa-routing-indicator global-title version 10.8.0 dscp-mark de map-importance-level auto-
configure bssap-importance-level auto-configure cap-importance-level auto-configure map-
subsystem-number standard bssap-subsystem-number 192 cap-subsystem-number unused
Step # Action
version 10.8.0;
dscp-mark de;
map-importance-level auto-configure;
bssap-importance-level auto-configure;
cap-importance-level auto-configure;
map-subsystem-number standard;
bssap-subsystem-number 192;
cap-subsystem-number unused;
Step # Action
ugl-retry 4;
sai-retry 3;
afr-retry 2;
mc-fallback-timer 21;
sai-vector-limit 4;
odb-ps disable
user@host% set mme-sgsn interface ss7 remote-destination <destination point code> <remote-
subsystems (SSN)> xudt-option enable ludt-option disable
user@host% set mme-sgsn interface ss7 peer <peer name> app-server-name <app-server name> ip-
address1 <Peer IP Address> peer-type ipspSe association-creation initiate 0.0.0.0 <Peer Port>
standard 0
ALERT: For Gr, the SGSN is the SCTP Client. Set the association-creation to initiate.
Alert: For SIGTRAN m3ua ss7 peer configuration attribute 'association-creation' should always remain as 'initiate'
(the default).
Syntax:
user@host% set mme-sgsn interface ss7 dest-peer-map <destination point code> g-peer-name <g-peer
name>
Minimum result is two. One for outbound traffic and one for inbound traffic.
GTT Result for HLR01:
user@host% set mme-sgsn interface ss7 gtt-result 1 destination-point-code 100.90.200 routing-
indicator route-on-gt use-original-gtai enable subsystem 6 global-title-indicator gti2
translation-type 10 numbering-plan 1 nature-of-addr-indicator 4 global-title-addr-info
358203060000010
Step # Action
GTT Result for inbound traffic for SGSN MAP:
user@host% set mme-sgsn interface ss7 gtt-result 3 destination-point-code 5.200.3 routing-
indicator route-on-ssn use-original-gtai enable subsystem 149 global-title-indicator gti2
translation-type 10 numbering-plan 1 nature-of-addr-indicator 4 global-title-addr-info
01001001001
user@host% show mme-sgsn interface ss7 gtt-result 1
destination-point-code 100.90.200;
routing-indicator route-on-gt;
use-original-gtai enable;
subsystem 6;
global-title-indicator gti2;
translation-type 10;
numbering-plan 1;
nature-of-addr-indicator 4;
global-title-addr-info 358203060000010;
Step # Action
upper-gtai 0;
primary-result 2;
secondary-result 2;
global-title-indicator gti4;
translation-type 0;
numbering-plan 7;
nature-of-addr-indicator 4;
user@host% set mme-sgsn interface ss7 address-translation 24405 country-code 358 network-code 20
user@host% show mme-sgsn interface ss7 address-translation 24405
country-code 358;
network-code 20;
Please refer to the S6 configuration section under MME, as it is common to MME and SGSN. S6a
and S6d are the same interface for a node that is both MME and SGSN.
For a greater understanding of recent changes to GTP, see GTP Feature Overview.
Step # Action
1 Configure the GTP common profile parameters.
user@host% set mme-sgsn interface gtp max-pending-transactions 20000
user@host% set mme-sgsn interface gtp gtpv1-version 10.8.0
user@host% set mme-sgsn interface gtp gtpv2-version 10.10.0
Step # Action
1a Define the GTP default profile (Gn to GGSN, S4 to SGW).
The following ‘tunnel’ profile is shared for Gn/S4:
Syntax:
user@host% set mme-sgsn interface gtp profile default routing-instance <id> ip-address-v4 <Local
IPV4 address> ip-address-v6 <Local IPV6 address> dscp-mark <DSCP marking for gtp profile> echo-
interval <echo interval> t3-echo-response <echo response timer> n3-echo-request <number of echo
request> t3-response <response timer> n3-request <number of request> inactivity-timer <inactivity
timer> n3-request-session-creation 2 t3-response-session-creation 5 n3-request-bearer-
modification 2 t3-request-bearer-modification 5 n3-request-context-request 2 t3-response-context-
request 5 n3-request-forward-relocation 2 t3-response-forward-relocation 5
user@host> show mme-sgsn interface gtp profile default
routing-instance EPCCore01;
ip-address-v4 172.18.249.17;
ip-address-v6 ::;
dscp-mark de;
echo-interval 300;
t3-echo-response 15;
n3-echo-request 8;
t3-response 5;
n3-request 2;
inactivity-timer 10;
n3-request-session-creation 2;
t3-response-session-creation 5;
n3-request-bearer-modification 2;
t3-request-bearer-modification 5;
n3-request-context-request 2;
t3-response-context-request 5;
n3-request-forward-relocation 2;
t3-response-forward-relocation 5;
Optional: If you wish to separate the MME/SGSN source address for S4 and S11 paths, add another gtp-profile
specific to that interface and link it under the interface. This will override the default "tunnel" profile for that selected
interface.
user@host% set mme-sgsn interface gtp profile s4 ip-address-v4 5.6.7.8 routing-instance EPCCore01
user@host% set mme-sgsn interface s4 s4-if gtp-profile s4
2 Define the GTP mobility profile (Gp to SGSN, S3 to MME and S16 to SGSN).
Syntax:
user@host% set mme-sgsn interface gtp profile mobility routing-instance <id> ip-address-v4 <Local
IPV4 address> ip-address-v6 <Local IPV6 address> dscp-mark <DSCP marking for gtp profile> echo-
interval <echo interval> t3-echo-response <echo response timer> n3-echo-request <number of echo
request> t3-response <response timer> n3-request <number of request> inactivity-timer <inactivity
timer> n3-request-session-creation 2 t3-response-session-creation 5 n3-request-bearer-
modification 2 t3-request-bearer-modification 5 n3-request-context-request 2 t3-response-context-
request 5 n3-request-forward-relocation 2 t3-response-forward-relocation 5
Unless specified the Tunnel and Mobility IPs can be the same.
Step # Action
t3-response-session-creation 5;
n3-request-bearer-modification 2;
t3-request-bearer-modification 5;
n3-request-context-request 2;
t3-response-context-request 5;
n3-request-forward-relocation 2;
t3-response-forward-relocation 5;
3 S4-SGSN only: Enable the S4 feature, and select treatment for non-epc UE and data plane usage.
Step # Action
name ggsndt1 direct-tunnel enable blacklisted false;
Step # Action
1 Create the SGSN Pool.
The UMTS interface to the radio access equipment (e.g. RNC then BTS), runs over the IP transport
mechanism. Legacy ATM transport is not directly supported by the SGSN. In order to ensure
endpoint redundancy, each RNC should be configured to terminate to two different IPSP instances on
the SGSN, which reside on different IPSP-instances. Additionally, path redundancy is support by
enhancing the IPSP instance to have multi-homed local endpoints. This allows upgrade and in-
service patching procedures to occur and still maintain the Iu connection to each RNC.
Step # Action
1 Configure the SCTP layer.
user@host% set mme-sgsn interface sctp sctpIu rto-initial 500 rto-min 100 rto-max 1000 path-max-
retrans 2 association-init-max-transmit 2 association-max-retrans 2 no-delay true max-out-stream
10 max-in-stream 10 hb-interval 30 rx-buffer-size 149000 tx-buffer-size 4900 sack-delay 200 sack-
frequency 2
3 Create the Iu-If interface (singleton) with details, and optionally enable HSPA+ feature.
For an S4-SGSN, the Iu standards version must be 9.5.0 or greater.
Determine if the SS7 networking will be ANSI or ITU (this affects the point code formats) and select the
signaling-variant here.
Step # Action
reset-timer-trafr 90;
reset-retry-trafr 1;
sccp-class-of-service connection-less,connection-oriented;
sccp-connection-establish-timer 60;
sccp-inactivity-test-timer 300;
sccp-inactivity-receive-timer 660;
sccp-timer-release 10;
sccp-timer-repeat-release 20;
sccp-pause-delay-timer 120;
trab-assgt-timer 4;
trab-assgt-retry 4;
reloc-alloc-timer 10;
reloc-complete-timer 25;
iu-release-timer 5;
iu-signaling-act-timer disable;
security-mode-timer 5;
security-mode-retry 2;
srns-rau-timer 5;
inter-rat-indirect-forwarding never;
indirect-data-fwd-ip-address 0.0.0.0;
indirect-tunnel-hold-timer 5;
version 9.5.0;
reloc-resource-release-timer 5;
dscp-mark de;
signaling-variant itu;
To add support for the Enhanced SRNS (Serving Radio Network Subsystem) Relocation Procedures as well as
increasing the number of RNCs (Radio Network Controller) supported by the SGSN from 256 to 4096, enable the
hspa-plus feature.
4 Configure the Local Iu Server. A loopback interface is required for the Iu IPSP address.
user@host% set mme-sgsn interface iu ipsp-instance <unit> 0 ip-address1 <Access IP Address1> ip-
address2 <Access IP Address2> routing-instance1 <routing-instance1> routing-instance2 <routing-
instance2> port <Local Port>
user@host# show mme-sgsn interface iu ipsp-instance
0 ip-address1 172.18.249.2 ip-address2 172.18.249.4 routing-instance1 EPCAccess01 routing-
instance2 EPCAccess02 port 3905
As with the Sigtran ASP each instance of the IPSP service unit require a unique ipsp-instance and IP address.
5 Configure Iu M3UA Pointcode.
user@host% set mme-sgsn interface iu m3ua m3ua-profile m3uaIu <local point code> 2.200.3
Step # Action
ALERT: The SGSN will need to be engineered to properly distribute the RNCs across the configured IU Service
Units.
user@host% set mme-sgsn interface iu peer AIRVANAFAP01 port <number> ip-address1 <M3UA peer
process1 ip address> ip-address2 <M3UA peer process2 ip address> peer-type <type of m3ua process>
compat-mode <compatible m3ua mode> association-creation <sctp association to peer> network-
appearance <network appearance>
user@host% show mme-sgsn interface iu peer AIRVANAFAP01
port 2905;
ip-address1 172.16.239.205;
ip-address2 0.0.0.0;
peer-type ipspSe;
compat-mode standard;
association-creation initiate;
network-appearance unused;
The Gb interface to the GPRS radio access equipment (BSS, i.e. PCU then BTS) supports IP
(Internet Protocol) or FR (Frame Relay) transport, known as Gb/IP and Gb/FR respectively. Either
Gb/IP or Gb/FR transport is specified when an NSE is provisioned. The maximum number of NSE’s
per SGSN is 600.
For Gb/IP, an NSE may be configured in dynamic mode or static mode, the former being preferable.
For Gb/FR, external equipment for E1/T1 Frame Relay conversion to Ethernet is provided by a
DragonWave Pseudowire Multi-Service Packet Node (PMSPN) Fusion A20.
To support Gb NSE redundancy, an “nsvl-group” is created with two “nsvl” instances, where each
“nsvl” instance is linked to a unique DATA VM. There may be multiple “nsvl-group” instances
depending on the number of NSEs that the SGSN needs to support. Up to 100 NSE’s may be linked
to single nsvl-group.
The following example creates one Gb “nsvl-group” with two “nsvl-group nsvl” instances, each linked
to a unique DATA VM, for redundancy. This is required to avoid Gb outages during in-service-
patching and upgrade.
Step # Action
1 Add Gb interface (singleton) to the SGSN. If the Gb/FR option is used, then more parameters are
required to specify local VR and IP. Gb/FR and Gb/IP NSEs can be supported concurrently on an
SGSN, but must use unique IP addresses. The version must be 9 or greater for an S4-SGSN.
short form for Gb IP only system:
user@host% set mme-sgsn interface gb gb-if fr-local-ip-address 0.0.0.0 payload-compression-
override disable rim-addr-err-pdu-status-cause invalid-mandatory-information ms-flow-control-th-
timer disable cancel-th-on-ms-move-to-standby disable alternative-bmax 0 version 9 dscp-mark de
user@host% show mme-sgsn interface gb gb-if
fr-local-ip-address 0.0.0.0;
payload-compression-override disable;
rim-addr-err-pdu-status-cause invalid-mandatory-information;
ms-flow-control-th-timer disable;
cancel-th-on-ms-move-to-standby disable;
alternative-bmax 0;
version 9;
dscp-mark de;
2 Create redundant Gb “nsvl-group” instances for each pair of DATA VM’s – Repeat as needed.
ALERT: For Gb/IP and depending upon PCU implementation, preferably all ‘nsvl-group nsvl’ IP/Port
addresses must be provisioned at each PCU to support ‘dynamic configuration’ of the Gb interface.
However, at minimum one ‘nsvl-group nsvl’ IP/Port address must be provisioned at each PCU.
For Gb/IP, the SGSN also supports ‘static configuration’ over the Gb interface, although ‘dynamic
configuration’ is preferred.
If only Gb/FR is to be provisioned, then routing-instance, ip-address, and port is optional and should not
be specified; if both Gb/FR and Gb/IP is to be provisioned on the service-units then vr-id, ip-address,
and port must be configured.
Create “nsvl-group x nsvl 0” and associate it with a DATA VM:
user@host% set mme-sgsn interface gb nsvl-group <#> nsvl 0 unit-id <#> local-ip-address <ip
address> local-port <#> routing-instance <string>
Create “nsvl-group x nsvl 1” and associate it with its redundant DATA VM:
user@host% set mme-sgsn interface gb nsvl-group <#> nsvl 1 unit-id <#> local-ip-address <ip
address> local-port <#> routing-instance <string>
3 Create a NSE (Network Service Entity) and link it to an “nsvl-group”. The NSE will be redundant since it
will exist on the two DATA VM’s that are associated with “nsvl-group <#> nsvl 0” and “nsvl-group <#>
nsvl 1” above. Repeat as needed.
Syntax:
user@host% set mme-sgsn interface gb nse <nse-id> nsvl-group <#> transport-type <transport-type>
port <local-udp-port>
for IP based PCU:
user@host% set mme-sgsn interface gb nse 12 nsvl-group <#> transport-type ip port 701
Step # Action
4 Setting: Display the default BSSGP configuration and modify if desired.
The default values are shown below.
Step # Action
1 Optional GbIP only: Add static NSE config if PCU does not support dynamic mode.
Syntax:
user@host% set mme-sgsn interface gb static-nse-remote-endpoint <remote-ip-address> port <port>
nse-id <nse-id> signaling-weight <signaling-weight>
2 Optional GbIP only: Display the NS (Network Service) layer of the Gb Network Service default values
and adjust if desired.
user@host% set mme-sgsn interface gb ns sns-config-retry 3 sns-delete-retry 3 sns-guard-timer 3
ns-test-timer 30 ns-alive-timer 3 ns-alive-retry 10
user@host% show mme-sgsn interface gb ns
sns-config-retry 3;
sns-delete-retry 3;
sns-guard-timer 3;
ns-test-timer 30;
ns-alive-timer 3;
ns-alive-retry 10;
Step # Action
1 Create a frame relay tunnel.
Repeat the command below to create as many fr-tunnels as needed to the applicable remote
DragonWave PMSPN TLC IO cards.
Step # Action
Syntax:
user@host% set mme-sgsn interface gb fr-tunnel <name> remote-ip-address <remote-ip-address>
2 Repeat the command below to create as many Gb/FR NSVCs as needed for the NSE(s) created in the
previous steps. Gb/FR NSVCs must be administratively assigned to a particular “nsvl-group nsvl”
instance associated with the NSE, and they must be split (e.g. if an Gb/FR NSE has four NSVCs, then
two NSVCs should be assigned to “nsvl-group x nsvl 0”, and 2 NSVCs should be assigned to “nsvl-
group x nsvl 1”).
Syntax:
user@host% set mme-sgsn interface gb nsvc-fr <nse-id> nsvc-id <nsvc-id> dlci <dlci> fr-tunnel
<fr-tunnel> remote-port <remote-port> nsvl-group <x> nsvl <0 or 1>
user@host% set mme-sgsn interface gb nsvc-fr 200 nsvc-id 2000 dlci 20 fr-tunnel fr1 remote-port
2500 nsvl-group <x> nsvl <0 or 1>
Step # Action
1 Configure the routing areas and how they map to access nodes.
If the rnc list or nse list is empty use ‘none’.
Syntax:
user@host% set mme-sgsn service-area routing-area <plmn> <lac> <rac> rncid-list <rncid-list>
nseid-list <nseid-list> timezone <system timezone>
umts example:
user@host% set mme-sgsn service-area routing-area 50502 0x2280 0x25 nseid-list none rnc-id list
602 timezone <system timezone>
This presumes that the Gr interface is already configured and that the SCTP and M3UA layers are
reused for Gf.
Step # Action
1 Add the Gf Interface. A loopback interface is required for the Gf ASP IP address.
2 If the SGSN will connect directly to the EIR, add the Peer Configuration.
Syntax:
user@host% set mme-sgsn interface ss7 remote-destination destination-point-code <destination
point code> xudt-option <xudt-option> ludt-option <ludt-option> remote-subsystems <remote
subsystem number (SSN)>
user@host% show mme-sgsn interface ss7 remote-destination destination-point-code 100.91.200
user@host% set mme-sgsn interface ss7 peer <g-peer name> app-server-name <app-server name> peer-
type ipspSe association-creation initiate ip-address1 172.16.241.190 ip-address2 0.0.0.0 port
<Peer Port> compat-mode standard network-appearance 0
Step # Action
}
ALERT: For Gf the SGSN is the SCTP Client. Set the association-creation to initiate.
Alert: For SIGTRAN m3ua ss7 peer configuration attribute 'association-creation' should always remain as 'initiate'
(the default).
Syntax:
user@host% set mme-sgsn interface ss7 dest-peer-map <destination point code> <g-peer name>
The minimum result is two. One for outbound traffic and one for inbound traffic.
GTT Result for EIR01:
user@host% set mme-sgsn interface ss7 gtt-result 10 destination-point-code 100.91.200 routing-
indicator route-on-gt use-original-gtai enable global-title-indicator gti2 translation-type 10
numbering-plan 1 nature-of-addr-indicator 4 global-title-addr-info 4691129999
Step # Action
Please refer to the S13 configuration section under MME, as it is common to MME and SGSN. If S13
is configured and Gf if configured, an S4-SGSN will only use S13.
Step # Action
1 Enable the ‘ge’ feature.
user@host% set mme-sgsn feature control ge status on
1a Add the Ge Interface. A loopback interface is required for the Ge ASP IP address.
Step # Action
2 Configure TCAP engineering parameters.
user@host% set mme-sgsn interface ss7 peer g-peer-name PP2 app-server-name default ip-address1
172.16.241.148 peer-type ipspSe association-creation initiate ip-address2 0.0.0.0 port 41730
compat-mode standard network-appearance 0
Alert: For SIGTRAN m3ua ss7 peer configuration attribute 'association-creation' should always remain as
'initiate' (the default).
user@host% set mme-sgsn interface ss7 dest-peer-map <destination-point-code> 100.94.200 <g-peer-
name> PP2
user@host% set mme-sgsn interface ss7 gtt-entry 7 lower-gtai 4694 upper-gtai 0 primary-result 6
secondary-result 6 global-title-indicator gti2 translation-type 10 numbering-plan 1 nature-of-
addr-indicator 4
5.2.4 Gd - SMS
Add one or more of the following:
The following may have been added earlier (see Gr section above):
user@host% set mme-sgsn interface ss7 gtt-entry <instance> lower-gtai <lower global title
address info> upper-gtai <upper global title address info> primary-result <primary
result> secondary-result <secondary result> global-title-indicator
<gti0/gti1/gti2/gti3/gti4> translation-type <translation type> numbering-plan <numbering
plan> nature-of-addr-indicator <nature of address indicator>
Step # Action
1 Add Gd Interface and configure CP Response Retries. A loopback interface is required for the Gd ASP
IP address.
user@host% set mme-sgsn interface gd gd-if cp-response-retry 1
2 Configure SMS-C.
3 If the SGSN will connect directly to the SMS-C, add the Peer Configuration.
Syntax:
user@host% set mme-sgsn interface ss7 remote-destination <destination point code> ludt-option
disable xudt-option disable remote-subsystems <remote subsystem number (SSN)>
Syntax:
user@host% set mme-sgsn interface ss7 peer <g-peer name> app-server-name <app-server name> peer-
type ipspSe association-creation initiate ip-address1 172.16.241.190 <ip-address2> 0.0.0.0 port
<Peer Port> compat-mode standard network-appearance 0
ALERT: For Gd the SGSN is the SCTP Client. Set the association-creation to initiate.
Alert: For SIGTRAN m3ua ss7 peer configuration attribute 'association-creation' should always remain as 'initiate'
(the default).
Syntax:
user@host% set mme-sgsn interface ss7 dest-peer-map add <destination point code> g-peer-name <g-
peer name>
Step # Action
The minimum result is two. One for outbound traffic and one for inbound traffic.
GTT Result:
user@host% set mme-sgsn interface ss7 gtt-result 11 destination-point-code 100.93.200 routing-
indicator route-on-gt use-original-gtai enable subsystem unused-or-default global-title-indicator
gti2 translation-type 10 numbering-plan 1 nature-of-addr-indicator 4 global-title-addr-info
4691139999
user@host% show mme-sgsn interface ss7 gtt-result
instance 11;
destination-point-code 100.93.200;
routing-indicator route-on-gt;
use-original-gtai enable;
subsystem 8;
global-title-indicator gti2;
translation-type 10;
numbering-plan 1;
nature-of-addr-indicator 4;
global-title-addr-info 4691139999;
Step # Action
1 Enable the feature and add the Gs Interface. A loopback interface is required for the Gs ASP IP
address.
Step # Action
Syntax:
user@host% set mme-sgsn interface gs gs-if global-title <global-title> cgpa-routing-indicator
<cgpa-routing-indicator> nature-of-address <nature-of-address> return-option <return-option>
xudt-option <xudt-option> registration-request-timer <registration-request-timer> loc-upd-on-lai-
mismatch <loc-upd-on-lai-mismatch> loc-upd-timer <loc-upd-timer> explicit-gprs-detach-timer
<explicit-gprs-detach-timer> explicit-gprs-detach-retry <explicit-gprs-detach-retry> explicit-
imsi-detach-timer <explicit-imsi-detach-timer> explicit-imsi-detach-retry <explicit-imsi-detach-
retry> implicit-imsi-detach-timer <implicit-imsi-detach-timer> implicit-imsi-detach-retry
<implicit-imsi-detach-retry> reset-timer <reset-timer> version <version>
Step # Action
user@host% set mme-sgsn service-area vlr-pool <vlr-pool name> vlr-selection-digits <vlr-
selection-digits> vlr-name <vlr name>
4 If the SGSN will connect directly to the VLR, add the Peer Configuration.
Syntax:
user@host% set mme-sgsn interface ss7 remote-destination <destination point code> xudt-option
disable ludt-option disable remote-subsystems <remote subsystem number (SSN)>
ALERT: For Gs, the SGSN is the SCTP Client. Set the association-creation to initiate.
Alert: For SIGTRAN m3ua ss7 peer configuration attribute 'association-creation' should always remain as 'initiate'
(the default).
Syntax:
user@host% set mme-sgsn interface ss7 dest-peer-map <destination point code> g-peer-name <g-peer
name>
user@host% show mme-sgsn interface ss7 dest-peer-map
100.92.200 g-peer-name VLR01;
The minimum result is two. One for outbound traffic and one for inbound traffic.
GTT Result for VLR01:
user@host% set mme-sgsn interface ss7 gtt-result 12 destination-point-code 100.92.200 routing-
indicator route-on-gt use-original-gtai enable subsystem unused-or-default global-title-indicator
gti2 translation-type 10 numbering-plan 1 nature-of-addr-info 4 global-title-addr-info 4691139999
user@host% show mme-sgsn interface ss7 gtt-result
Step # Action
12 {
destination-point-code 100.92.200;
routing-indicator route-on-gt;
use-original-gtai enable;
subsystem 7;
global-title-indicator gti2;
translation-type 10;
numbering-plan 1;
nature-of-addr-indicator 4;
global-title-addr-info 4691149999;
}
Step # Action
1 Configure SGSN accounting parameters (Ga profile).
Step # Action
user@host% set mme-sgsn interface ga sgsn-ga-profile version <9.6.1/10.12.0>
3 Create the Ga Interface. A loopback interface is required for the Ga interface IP address.
Step # Action
1 Define the CGF Server (Primary/Secondary).
Syntax:
user@host% set mme-sgsn interface ga cgf <CGF IP Address> port <CGF Port> configured-role
<primary/secondary>
The following procedure describes how to pull billing records (CDRs) that have been stored to disk in
3GPP TS 32.297 compliant format from the SGSN.
Do not pull the files from the SGSN’s disk if a CGF is provisioned on the SGSN. In that case, the files
are automatically read from the disk and sent to the CGF without any operator intervention required. If
files are pulled in this case, it could lead to duplicate records and unexpected behavior on the SGSN.
This procedure only applies to the case when a CGF is not provisioned on the SGSN. In that case,
the files are written to disk, but the operator must periodically SFTP the files off of the SGSN to
ensure the disk partition does not become full.
The frequency with which the files should be pulled is dependent on the amount of traffic and
accounting provisioning. A Disk High Usage alarm (GaBilling1300) is generated when the billing
partition reaches specific thresholds (Minor: 30%, Major: 50%, Critical: 90%, and Clear: 25%). Use
this alarm as a guide to when the files should be pulled, but do not wait for the disk to become full
since CDRs will be discarded if that occurs.
Step # Action
1 Check the current CGF provisioning.
Step # Action
1 SFTP to the SGSN (non-gui sftp client is shown but any sftp client should work).
sftp sftpuser@vmme20
sftp> cd /usr/local/epc/accounting/closed
sftp> ls -l
-rw-r--r-- 1 root root 61365 Mar 14 17:45 vmme20_-_1.20120314_-
_1745+0000.mcdr_v36_ber.closed
-rw-r--r-- 1 root root 1048411 Mar 14 17:46 vmme20_-_2.20120314_-
Step # Action
_1745+0000.scdr_v36_ber.closed
-rw-r--r-- 1 root root 1048423 Mar 14 17:47 vmme20_-_3.20120314_-
_1746+0000.scdr_v36_ber.closed
-rw-r--r-- 1 root root 884693 Mar 14 17:48 vmme20_-_4.20120314_-
_1748+0000.scdr_v36_ber.closed
-rw-r--r-- 1 root root 152188 Mar 14 17:48 vmme20_-_5.20120314_-
_1748+0000.scdr_v36_ber.closed
4 Make a new directory to move the files to pull into (a name besides ‘pull’ can be used).
5 Move (rename) the ‘.closed’ files into the new ‘pull’ directory.
ALERT: Use the above command for each file to be pulled. A graphical sftp client will make this much easier since
multiple files can be highlighted and dropped into the ‘pull’ folder.
ALERT: A new directory is created because the SGSN may still be writing new files to disk while the files are being
pulled. This allows for an easy way to tell which files should be deleted after the pulling is complete. Some graphical
clients allow you to ‘move’ files (copy and then delete) directly from remote to local and thus the need to move them to
a temporary directory is no longer necessary.
6 Pull the files from the ‘pull’ directory.
sftp> cd pull
sftp> get *
Fetching /usr/local/epc/accounting/closed/pull/vmme20_-_1.20120314_-_1745+0000.mcdr_v36_ber.closed
to vmme20_-_1.20120314_-_1745+0000.mcdr_v36_ber.closed
/usr/local/epc/accounting/closed/pull 100% 60KB 59.9KB/s 00:00
Fetching /usr/local/epc/accounting/closed/pull/ vmme20_-_2.20120314_-
_1745+0000.scdr_v36_ber.closed to vmme20_-_2.20120314_-_1745+0000.scdr_v36_ber.closed
/usr/local/epc/accounting/closed/pull 100% 1024KB 1.0MB/s 00:00
sftp> lls -l
-rw-r--r-- 1 root root 61365 Mar 14 17:45 vmme20_-_1.20120314_-
_1745+0000.mcdr_v36_ber.closed
-rw-r--r-- 1 root root 1048411 Mar 14 17:46 vmme20_-_2.20120314_-
_1745+0000.scdr_v36_ber.closed
ALERT: The command is ‘lls’ not ‘ls’. ‘lls’ is for ‘local ls’.
ALERT: It is probably better to check the destination directory using a local file manager or local terminal.
8 Clean up.
sftp> rm *
sftp> cd ..
sftp> rmdir pull
5.2.7 Gn S4 Selection
This feature enhances the S4-SGSN node to provide the following functionality for UEs in Iu mode:
gateway interface selection based on configurable options (Gn, S4 preferred or Gn/S4 using EPC
capability)
Step # Action
1 Verify the S4-SGSN feature is enabled.
[edit]
user@host% show mme-sgsn feature control | tab
NAME STATUS
--------------------------
direct-tunnel off
equip-id-check off
fxa off
gs on
hspa-plus on
ims-emergency on
lawful-intercept on
s101 off
s102 on
s4-sgsn on
sbc off
sgs on
trace on
ue-purge off
[ok][2014-09-03 10:55:37]
Use the set command (set mme-sgsn feature control s4-sgsn status on) if the feature is off.
2 Set the dns-preference attribute to the preferred value. In order to support both DNS s-naptr and classic
query procedures, the value unknown should be configured.
[edit]
user@host% set mme-sgsn subscriber subscriber-plmn 31298 dns-preference unknown
[ok][2014-09-03 10:55:37]
[edit]
user@host% set mme-sgsn subscriber subscriber-class default gateway-interface-sel ue-epc-
capability
[ok][2014-09-03 10:55:37
Step # Action
1 Add the S4 interface table to control any optional features for that interface, or to point to a unique gtp-
profile for this particular interface when required.
The uli-inclusion parameter controls if the SGSN will send ULI in the below GTPv2 messages if it is set to the enum
value 'force-in-all':
• Create Session Request(HO/TAU)
• Modify Bearer Request
• Delete Session Request
• Delete Bearer Command
The ue-timezone-inclusion parameter controls if the SGSN will send UE-Timezone in GTPv2 messages.
In this example, a separate gtp profile is used for the s4 interface.
Step # Action
1 Create an SGSN pool.
Step # Action
1 Enable the Traffic Shaping feature.
When enabled, the SGSN will use the allocation and retention priority and traffic class from the QoS of a mobile to
retrieve the DSCP Mark for inclusion in data packets that pass through this node.
2 Modify the Differentiated Service (DS) Code Point (DSCP) value for signalling traffic and the allocation
and retention priority assigned to a mobile that is a roamer.
3 Modify the TrafficClass (Tc) and/or Traffic Handling Priority (THP) provisioning data of the SGSN.
user@host% set mme-sgsn feature traffic-shaping profile priority 3 traffic-class streaming data-
dscp-mark af13 preemption-capability auto-configure preemption-vulnerable enabled qa auto-
configure
Step # Action
2 traffic-class background {
data-dscp-mark af12;
preemption-capability disabled;
preemption-vulnerable auto-configure;
queuing-allowed disabled;
}
3 traffic-class streaming {
data-dscp-mark af13;
preemption-capability auto-configure;
preemption-vulnerable enabled;
queuing-allowed auto-configure;
}
1 traffic-class interactive-thp1 {
data-dscp-mark af21;
preemption-capability auto-configure;
preemption-vulnerable enabled;
queuing-allowed auto-configure;
}
2 traffic-class interactive-thp2 {
data-dscp-mark af22;
preemption-capability enabled;
preemption-vulnerable disabled;
queuing-allowed enabled;
}
3 traffic-class interactive-thp3 {
data-dscp-mark af23;
preemption-capability disabled;
preemption-vulnerable auto-configure;
queuing-allowed disabled;
}
Step # Action
1 Activate the RAN information management feature.
user@host% set mme-sgsn feature control ran-info-management status on
Step # Action
1 Activate the RAN admission control feature.
configured and default value being ‘no-detach-msg’. Operators can also configure a list of a
maximum of 40 APN-NI`s, which will exclude the subscribers having those APN-NI from getting
detached and default being ‘none’.
Step # Action
1 Turn-On the feature.
user@host% set mme-sgsn feature control sgsn-sessionless-sub-detach status on
The feature entity ‘sgsn-sessionless-sub-detach’ is populated with default values of each attribute only after the
feature is turned on. Initially, it will be empty with no values.
2 Check the default values in sgsn-sessionless-sub-detach feature entity.
Step # Action
1 Activate Packet-Flow-Control Feature.
Syntax:
user@host% set mme-sgsn interface gb packet-flow-management packet-flow-context <packet-flow-
context> pfc-flow-control <pfc-flow-control> pfc-flow-control-tf-timer <pfc-flow-control-tf-
timer> average-pfcs-per-mobile <average-pfcs-per-mobile> t7-timer <t7-timer> t7-retries <t7-
retries> packet-flow-timer <packet-flow-timer>
ALERT: Default entries for the attributes are shown above. All the attributes can be configured as per their need.
Step # Action
1 Set the imei check lifetime and imei-check-during-internode.
Step # Action
user@host% set mme-sgsn feature equip-identity-check allow-gray-mobile enable log-black-grey-
mobile disable imei-check-lifetime enable imei-check-during-internode off
During IRAU, packets need to be buffered and then sent to the new SGSN when it is ready to service
the mobile. If the old SGSN is a 2G node, IRAU buffering always takes place on the SGSN. For 3G,
this buffering normally takes place on the RNC; however, if the RNC does not support buffering or if
the operator wants buffering to take place on the SGSN as an optimization, the following attribute can
be set.
Step # Action
1 Enable buffering on the old 3G SGSN.
user@host% set mme-sgsn interface sm old-sgsn-buffering enable
Note that provisionable parameters for IRAU buffering (irau-buff-max-bytes-per-session) are for
buffering on the new SGSN. For IRAU buffering on the old SGSN, provisionable parameters for
downlink buffering are used (dl-buff-max-bytes-per-session; again).
The SGSN supports RFC1144 IP header compression and V.42 bis user data compression over the
Gb interface. These can be used to decrease overhead and optimize radio resource usage efficiency
for 2G networks.
rfc1144-slots: the number of state slots that can be used by the RFC1144 algorithm
max-v42bis-entities: the number of sessions that can run data compression simultaneously as a
percentage of sd-capacity subs-per-su
v42bis-compression-direction: which data to compress – data from the mobile, data to the
mobile, or both
The common (header and data) compression parameter has the following meaning.
Step # Action
1 Configure IP header compression as per engineering requirements.
Syntax:
user@host% set mme-sgsn interface gb sndcp allowable-compression-alg <none | rfc1144> max-header-
compression-entities <value> rfc1144-slots <value>
PDU buffering is used to improve network performance by temporarily storing packets that cannot be
immediately transmitted. For 2G systems, this can be because the BSS has initiated flow control, LLC
is suspended, or the old SGSN is waiting for a context acknowledgement from the new SGSN during
IRAU. Packets are also buffered if LLC is operating in acknowledged mode. For 3G systems,
buffering occurs during paging and, if enabled, on the old SGSN during IRAU. Buffering for each of
these conditions is configured with the dl (downlink) buffer parameters shown below. In contrast, the
irau buffer parameters are used for buffering on the new SGSN during IRAU following the context
acknowledgement but before the lower layers are available.
<dl | irau>-buff-total-bytes-reserved: The maximum number of bytes that can be used for
buffering.
<dl | irau>-buff-max-pdu-per-session: The maximum number of PDUs that can be buffered per
session.
buff-check-interval: The periodic time interval (in 10ths of a second) after which the SGSN
attempts to transmit buffered PDUs.
pdu-lifetime-in-buff: The maximum amount of time (in seconds) that a single PDU can be
buffered.
percent-umts-buff: The percentage of buffering resources available for 3G data; the rest are
available for 2G data.
Step # Action
1 Set downlink buffering parameters as per engineering requirements.
Syntax:
user@host% set mme-sgsn engineering sd-engineering dl-percent-buffered-sessions <value> dl-buff-
max-pdu-per-session <value> dl-buff-max-bytes-per-session <value> dl-buff-total-bytes-reserved
<value>
The LLC layer of the GPRS dataplane configures with usable default values. The following examples
show the default values and the method to adjust if desired, but these steps are optional.
Step # Action
1 Optional: Display the LLC default configuration and modify if desired.
Step # Action
2 Optional: Display the LLC SAPI default configuration and modify if desired.
Activation of this feature may impact messaging and RF paging resources due to the fact that the PS
paging is delivered to all the cells in the RAI. Therefore, an engineering exercise is advised to
properly set the provisioning attribute.
Step # Action
1 Enable the Periodic Paging Resumption Feature.
use-pgwid-from-hss enable;
gw-reselection-retries 2;
Step # Action
1 Enabling the qos-negotiation attribute.
user@host% set mme-sgsn interface nas sm qos-negotiation enable
Step # Action
1 Modify and display the default QOS conversion details.
Step # Action
user@host% show mme-sgsn subscriber qos-conversion streaming
max-sdu-size 1500;
residual-ber residualber-5hundredth;
delivery-of-erroneous-sdus err-sdus-not-delivered;
arp-high 5;
arp-medium 10;
Step # Action
1 Enable the Second authentication attempts.
Step # Action
1 Setting of GMM cause code.
user@host% set mme-sgsn interface nas gmm-cause-mapping access-restriction-data cause-code no-
suitable-cell-in-la-15 application-unsupported cause-code no-suitable-cells-in-ra-15 command-
unsupported cause-code no-suitable-cells-in-ra-15 out-of-space cause-code congestion-22 too-busy
cause-code congestion-22 unable to comply cause-code no-suitable-cells-in-ra-15 unknown-serving-
node cause-code imsi-unknown-in-hss-2 roaming-restricted-in-mme-due-to-unsupported-feature cause-
code no-suitable-cells-in-la-15
Step # Action
1 Configure the capacity limit for each SC service unit.
This configures the maximum number of subscribers supported per SC application in this CALLP VM in
units of thousands.
Syntax:
Step # Action
user@host% mme-sgsn engineering sc-capacity <callp-service-id> subs-per-su <subs-per-su>
3 Configure the capacity limit for each SD service unit in thousands of subscribers.
Step # Action
critical-action ignore;
The SS7 congestion control is enhanced under current system overload control architecture. With
this feature, the local SCTP buffer usages of ASP instances are also included in calculating the
congestion level of the system, and the external SCON messages are process as part of
congestion control of the system.
The capability to specify the subsystem numbers for each 3GPP subsystems configured on the
SGSN.
The SIGTRAN process (ASP) sheds traffic based on the local subsystems and the configured
importance levels in case of congestion/overload.
An option is provided to configure/control the reconnect timer for ASP to establish the association
with applicable peers.
Note:
Multi-homing is not supported on the SIGTRAN interfaces (Gfreds and Iu).
The following table shows the configuration tables associated with these enhancements.
Step # Action
1 Display the reconnect timer and adjust if need be.
Step # Action
cong-report-frequency 8;
reconnect-timer 10;
Importance levels can be one of 1 2 3 4 5 6, auto-configure; Importance level messages (1::6 for ITU, 1::3 for ANSI);
default for all is auto-configure.
Subsystem numbers can be 2..255, standard, unused. The default for map is standard. The default for bssap and cap
is unused.
Step # Action
1 Set the NAS common parameters.
Attention:
ALL LI configuration must be done via the lea user account or a user with lea
permissions.
Enabling the LI feature must be done via the admin user account or a user with
admin permissions.
Step # Action
1 From the admin user account (or a user with admin permissions), enable the feature.
2 From the lea user account (or a user with lea permissions (see Adding a User), add the local LI
Interface. A loopback interface is required for the LI interface IP address.
lea@host % set mme-sgsn interface x x-if routing-instance EPCLi01 ip-address 172.16.52.32 max-
active-targets 1000 version 10.4.0 dscp-mark de
ALERT: This presumes that the IP appearance on a virtual router is already done in the prior steps.
3 Add entry for LIG (via the lea user account or a user with lea permissions).
lea@host % set mme-sgsn interface x lig <LIG IP Address> <LIG Port>
OAM interface where command access is only available to LEA users over an SSH interface
define the nodes that receive IRI messages for activated targets
Step # Action
1 Generate a TCP profile, if desired, to configure the socket used for LI delivery of IRI messages.
As an ‘admin’ or ‘lea’ user in config mode:
lea@host% set mme-sgsn interface tcp li_df2_profile keep-alive-timer 600 keep-alive-interval 75
keep-alive-probe 10
lea@host% set mme-sgsn interface x x2-if df md_node1 ip-address 192.168.10.1 port 102
lea@host% set mme-sgsn interface x x2-if df md_node2 ip-address 192.168.10.2 port 102
4 Add a target.
Step # Action
As a ‘lea’ user in operational mode:
lea@host> request mme-sgsn interface x x1 target add imsi 12345678901 lawful-intercept-identifier
“case 123” primary-delivery-connection md_node1 secondary-delivery-connection md_node2
5 Remove a target.
As a ‘lea’ user in operational mode:
lea@host> request mme-sgsn interface x x1 target delete imei 123456789012345 lawful-intercept-
identifier “warrant 42”
Step # Action
1 Add a new subscriber-class record to be used for classifying a specific set of roamers (up to 16 are
allowed).
ALERT: The imsi mask is a prefix match, so a 4 digit mask implies a match on the first 4 digits, for
example.
user@host% set mme-sgsn subscriber subscriber-class partnerx subscriber-type roamer csfb true
2 Set the protocol type for PLMN (where PLMN-ID is MCC and MNC, 5 or 6 digits total).
Chose gtp or pmip for protocol type (latter for CDMA networks typically) and enable/disable the sending
of NOR messages from MME to HSS.
3 Add naptr-record (only for MME) and a-record to resolve the APNs for the roamers. Refer to the DNS
Service section.
Step # Action
charging-char-for-gateway-sel disable;
Step # Action
1 Turn on the Seamless National Roaming feature in the feature-control table.
user@host% set mme-sgsn =subscriber imsi-range 31278 service-area global subscriber-class roam-
restrict
Step # Action
user@host% set mme-sgsn subscriber imsi-range 41323 service-area global subscriber-class roam-as-
home
9 Turn on the Local EPS QoS Control feature in the feature-control table.
Following is example configuration of IMS QoS Profile not allowing QCI 5 for default bearer and not
allowing QCI 1 for dedicated bearers:
Step # Action
1 Turn the feature on in the feature-control table.
2 Add a PLMN entry in the plmn-load-control entity and adjust settings per PLMN as desired.
5 Watch for PLMN based rate (MmeSc 1317) and level (MmeSc 1318) alarms.
Step # Action
Step # Action
1 Configure the behavior to use for service shutdown commands.
Perform offline software update of a node by moving all the UEs off of one MME/SGSN and to
other MME/SGSN in the same pool
Allows for targeted testing of new nodes, configuration, or functionality in the network by moving a
specified number of UEs without impact to the rest of the subscribers
The verb ‘move’ under ‘request mme-sgsn subscriber’ enables the operator to move a particular IMSI
from the source node (SGSN/MME) to the target node (SGSN/MME).
6.8 UE Purge
The optional UE purge feature is used by the MME to inform the HSS that it has deleted the
subscription and MM context of a detached UE.
The Operator can select MME’s behavior after the completion of the Detach procedure.
The Purge function can be triggered immediately or after a configured time period.
For each detach type and Subscriber type combination, an inactive age needs to be set. A purge for
an inactive UE meeting both the detach and UE type constraints of the UE purge setting will occur
only after its inactivity period exceeds the inactive age of the setting.
Setting the inactive age to immediate or a very low value (like less than a day) will increase signaling
to/from the HSS in the case where UEs reattach shortly after being purged. A low value is only to be
used if the node is operating close to a UE context limit and requires contexts to be freed quickly for
use by other active UEs.
A value of 24 hours is useful to clean up contexts from inactive UEs such as roamers. If context
resources are not a concern but signaling traffic to the HSS is a concern, a high value of upto 31 days
can be set.
MME provides the following PMs to track messaging for purge activity.
Step # Action
1 Activate the feature.
user@host% set mme-sgsn feature purge-trigger <detach type> all ue-type all inactive-age
immediate
Step # Action
Step # Action
1 Display the default subscriber overload configuration.
Step # Action
major-abate 75;
major-action reject-all-itau-irau;
major-ovld disable;
critical-onset 90;
critical-abate 85;
critical-action reject-bearer-setup;
critical-ovld disable;
Step # Action
1 Ensure that all the interfaces are provisioned for version 10.
Common:
user@host% set mme-sgsn interface gtp gtpv2-version 10.10.0
user@host% set mme-sgsn interface gtp gtpv2-version 10.10.0
user@host% set mme-sgsn interface gtp gtpv1-version 10.8.0
user@host% set mme-sgsn interface gtp gtpv1-version 10.8.0
user@host% set mme-sgsn interface nas nas-common version 10.10.0
MME/Combo:
user@host% set mme-sgsn interface s1 s1-if version 10.6.0
user@host% set mme-sgsn interface s6 s6-if version 10.7.0
SGSN/Combo:
user@host% set mme-sgsn interface iu iu-if version 10.9.0
user@host% set mme-sgsn interface ss7 ss7 version 10.8.0
3 Configure a new sctp-profile for FGW and link it to the s1 s1-if fgw-sctp-profile.
Add a new sctp-profile for FGWs. Use higher values for rx-buffer-size and tx-buffer-size than are used for the default
s1 sctp-profile (ALERT: the values below are just examples):
user@host% set mme-sgsn interface sctp fgwdefaultsctp rto-initial 500 rto-min 100 rto-max 1000
association-init-max-transmit 2 association-max-retrans 2 path-max-retrans 2 no-delay true max-
in-stream 10 max-out-stream 10 hb-interval 30 rx-buffer-size 1000000 tx-buffer-size 2000000 sack-
delay 200 sack-frequency 2
Step # Action
If spare capacity does not exist on the system then existing rnc entities must be removed to make space for the fgw-
rnc entries. The following action is not required on systems with spare capacity.
user@host% delete mme-sgsn interface iu rnc 123
Add fgw-rnc entry:
user@host% set mme-sgsn interface iu fgw-rnc 123 destination-point-code 2.42.3 callp-service-id 5
direct-tunnel enable max-iu-connections 1000
user@host% show mme-sgsn interface iu fgw-rnc
123 {
destination-point-code 2.42.3;
callp-service-id 5;
direct-tunnel enable;
max-iu-connections 1000;
}
Configuring fgw-enb:
If spare capacity does not exist on the system then existing s1 entities must be removed to make space for the fgw-
enb entities.
Add s1 fgw-enb entries for each FGW:
user@host% set mme-sgsn interface s1 fgw-enb 11111 enodeb-id 1234 ip-address1 172.16.241.148 ip-
address2 0.0.0.0 unit 4
Step # Action
1 Provision or ims-voice-sgsn for the nodal control for an MME or S4-SGSN.
Step # Action
MME:
Local EPS Qos Control is performed during Attach, PDN Connectivity, UE/GW Initiated Activation,
UE/GW/HSS Initiated Modification, Inter MME/Inter RAT Tracking Area Update and Inter MME/Inter
RAT S1 Handover procedures.
S4-SGSN:
Local EPS Qos Control is performed during the Session Management procedures of PDP Context
Activation, PDP Context Modification, HSS Initiated Modification, Inter-SGSN Routing Area Update,
and Inter-SGSN SRNS Handover (for 3G mobiles).
Step # Action
1 Configure a bearer-limits-profile for Local EPS Control default bearers (Note: Provision QCIs in the
range of 5-9).
user@host% set mme-sgsn subscriber bearer-limits-profile local-qos-default1 qos-class-identifier
qci5 arp-priority 1 pvi disable pci enable max-bitrate-uplink 0 max-bitrate-downlink 0
guaranteed-bitrate-uplink 0 guaranteed-bitrate-downlink 0
Step # Action
user@host% show mme-sgsn subscriber bearer-limits-profile local-qos-default1 qos-class-identifier
qci5
arp-priority 1;
max-bitrate-uplink 0;
max-bitrate-downlink 0;
guaranteed-bitrate-uplink 0;
guaranteed-bitrate-downlink 0;
pci enable;
pvi disable;
2 Configure a bearer-limits-profile for Local EPS dedicated bearers (Note: Provision QCIs in the range of
1-9).
Step # Action
local-eps-dedicated2 qos-class-identifier qci3
arp-priority 2;
pci enable;
pvi disable;
max-bitrate-uplink 200000;
max-bitrate-downlink 200000;
guaranteed-bitrate-uplink 250000;
guaranteed-bitrate-downlink 250000;
When the feature is enabled, the following are the NPLI sent by the MME/SGSN to the SGW:
Step # Action
1 Verify/set the gtp tunnel version for gtpv2 to Rel 10.
The cases where the MME/SGSN fails to select APN as per APN selection described in 3GPP TS
23.060 ANNEX A areas under:
APN requested, APN(R), by the User Equipment (UE), does not match any APN configuration in
the subscription and there is no wild card (WC) in the subscription
APN(R), does not match any APN configuration in the subscription but there is a wild card and
hence the APN(R) is used with the wild card but the DNS query fails
No APN(R) and there is more than one APN configuration in the subscription
This feature allows the operator to configure a choice for the APN selection in the above cases. The
operator could select an APN from the subscription if available, or configure to use a locally
configured APN.
Step # Action
1 Check the apn-selection-profile linked to the subscriber class for which the feature is to be turned on.
apn-selection-profile exampleapnselection;
}
a large number of UEs with similar cellular activity patterns, for instance UE activation during
morning and evening commutes. (Predictable)
a large number of UEs initiating access to the network due to specific events such as New Year’s
Eve. (Predictable)
a large number of UEs moving from one MME to another MME due to planned or unplanned
MME outage. (Both)
a large number of UEs initiating access to the network due to a natural event. (Unpredictable)
UPM load balancing functionality consists of UPM congestion monitoring and load balancing
management procedures. UPM congestion monitoring includes a new alarm that is raised and
cleared based on configurable thresholds, and CLI display enhancements.
Step # Action
1 Configure UPM TPS Alarm Congestion Threshold.
[ok][2015-01-21 10:34:20]
In previous functionality the MME load balanced the eNBs across the VMs when they initially
connected to it. This distribution of eNBs across VMs only changed if there was a VM SHUTDOWN, a
VM reboot or if the eNB re-connected. This prevented the MME from taking advantage of a newly
added capacity. The new eNB relocation functionality allows operators to move eNBs to a newly
added VM, or any other VM, without requiring the eNB to re-connect and thus allowing operators a
chance to manipulate the eNB distribution across the VMs.
The steps presented here are an example of how to use this feature to move eNB from one VM to
another.
Step # Action
1 Find where the eNB is currently located if the Global eNB id of the eNB to be relocated is known:
[ok][2015-01-21 08:32:11]
If the global enb id is not known then one of the below can be used as key:
• [enodeb-name <name>] or
• [ip-address1 <address> [ip-address2 <address>]])
If none of the above key(s) are known then the following command can be used to show all the eNBs connected to
the MME
user@host> show mme-sgsn interface s1 enb
PLMN ENODEB ENODEB IP ENODEB
ID ID TYPE IP ADDRESS1 ADDRESS2 NAME SERVICE
-------------------------------------------------------------------
504321 4032 macro 172.16.236.27 - eNB01 sig-1
504321 4034 macro 172.16.236.27 - eNB02 sig-0
504321 4036 macro 172.16.236.26 - eNB03 sig-0
504321 4038 macro 172.16.236.27 - eNB04 sig-0
504321 5039 macro 172.16.236.26 - HeNB01 sig-0
504321 7033 macro 172.16.241.185 - eNB03 sig-0
[ok][2015-01-21 08:34:47]
2 To display the load status of the target or all the SIG VMs use the following command:
Step # Action
callp0-1 system service=callp0/unit=1 - CpuUtilization 7
callp0-1 system service=callp0/unit=1 - MemoryUtilization 7
data-0 system service=data/unit=0 - CpuUtilization 1
data-0 system service=data/unit=0 - MemoryUtilization 5
data-1 system service=data/unit=1 - CpuUtilization 0
data-1 system service=data/unit=1 - MemoryUtilization 2
mgmt-0 system service=mgmt/unit=0 - CpuUtilization 1
mgmt-0 system service=mgmt/unit=0 - MemoryUtilization 4
mgmt-1 system service=mgmt/unit=1 - CpuUtilization 1
mgmt-1 system service=mgmt/unit=1 - MemoryUtilization 3
rm-0 system service=rm/unit=0 - CpuUtilization 1
rm-0 system service=rm/unit=0 - MemoryUtilization 4
rm-1 system service=rm/unit=1 - CpuUtilization 1
rm-1 system service=rm/unit=1 - MemoryUtilization 4
sig-0 system service=sig/unit=0 - CpuUtilization 1
sig-0 system service=sig/unit=0 - MemoryUtilization 4
sig-1 system service=sig/unit=1 - CpuUtilization 1
sig-1 system service=sig/unit=1 - MemoryUtilization 4
[ok][2015-01-21 08:43:44]
4 To learn the s1-capacity of each of the signaling VM, use the following command:
assuming that in our example the Global eNB id of the eNB is:
plmn-id : 50631
enodeb-id : 2213
enodeb-type : macro
and we want to move the eNB from sig-2 to sig-1 the command would be:
user@host> request mme-sgsn interface s1 enb relocate plmn-id 50631 enodeb-id 2213 enodeb-type
macro service sig-1
Per TAC
Per PLMN
IMS Voice
emergency support
timezone
Step # Action
1 Provision the IMS Emergency feature to on for the MME:
2 Provision location area(s) and associated VLR pool name(s) for the MME or S4-SGSN:
Syntax:
user@host% set mme-sgsn service-area location-area <plmn-id> <lac> vlr-pool-name <vlr pool name>
user@host% set mme-sgsn service-area location-area 50631 0x2280 vlr-pool-name vlrPool1
user@host% set mme-sgsn service-area location-area 50502 0x40ff vlr-pool-name vlrPool2
user@host% set mme-sgsn service-area tracking-area 000000 0x0000 lac 0x333 ims-support enable
emergency-support enable e-smlc-pool-name pool1
OR
user@host% set mme-sgsn interface nas gmm default ims-voice-sgsn on
4 Provision per-TAC support for IMS Voice over PS / IMS Emergency (MME):
user@host% set mme-sgsn service-area tracking-area 50631 0x1001 lac 0x333 ims-support disable
emergency-support disable
user@host% set mme-sgsn service-area tracking-area 50502 0x1001 lac 0x333 ims-support enable
emergency-support disable
5 Provision per-PLMN support for IMS Voice over PS for those PLMNs with roaming:
user@host% set mme-sgsn subscriber subscriber-plmn 50631 ims-support enable
Step # Action
user@host% set mme-sgsn subscriber subscriber-plmn 50631 ims-local-breakout allow
6.17.2 Timezone
Timezone provisioning may be performed independently of or in conjunction with the IMS Voice of
PS/IMS Emergency provisioning.
Step # Action
1 Provision the default timezone (MME):
user@host% set mme-sgsn service-area tracking-area 000000 0x0000 timezone America/Chicago
user@host% set mme-sgsn service-area routing-area 50631 0x1022 0x44 timezone America/Indianapolis
Step # Action
1 Unlock the different VMs to bring up the system.
2 Verify the status of the VMs and ensure that they are all unlocked, the default state for the VM.
[ok][2015-01-21 08:50:28]
Step # Action
1 Unlock the different VMs to bring up the system.
2 Verify the status of the VMs and ensure that they are all unlocked, the default state for the VM.
Step # Action
[ok][2015-01-21 08:50:28]
The MME/SGSN may operate as standalone MME, standalone S4-SGSN, standalone Gn-SGSN, or
a combined MME/SGSN. A combined node brings network level efficiency by reducing the number of
inter-nodal messages required to support the inter-RAT (Radio Access Technology) mobility events
when the UE moves between access technologies. And it brings nodal level efficiency by sharing
control plane resources between all three access technologies.
A GPRS/UMTS operator may choose to add LTE, and that can be done while the node is in-service
depending on the extent of re-engineering required. Or an LTE operator may want to add UMTS
support, which can be done while the node is in-service depending on the extent of re-engineering,
required. Before starting a procedure to add a new technology support to an existing node, consult
the engineering guidelines to ensure the proper number of VMs are added, and that the optimal
service topology is selected.
8.1.1 Overview
When adding MME function to an SGSN, there should already be compute nodes and VMs
configured for bearer path. However, the number may need to be adjusted if more control plane
capability is needed. If more VMs are needed for control plane, the existing DATA VMs should be
removed such that CALLP VMs can be configured. On the existing CALLP VMs, the subscriber
capacity and RNC fanout capacity may need to be reduced to leave room to introduce eNB fanout. In
the SGSN case, some impact to the UE may be required. The majority of the UEs can be migrated in-
service to other SGSNs in the same SGSN pool. This is done using the same CLI command
mentioned in the MME case (request mme-sgsn subscriber offload). If DATA VMs need to
be reconfigured for control plane, then some UEs will be required to re-activate their PDP contexts.
If a combined node is deployed from scratch, then the following interfaces should be deployed:
S1 interface
S11 interface
S6/S13 interface
S4 interface
S3/S16 interface
Gb interface
Other interfaces can be optionally deployed based on the need, these include:
Gr interface (for roamer from non-EPS capable networks, the UE can only acquire service on the
SGSN side)
Gn interface (if not already configured on the MME and mobility to legacy SGSN or session
management to legacy GGSN is required)
Gf interface (if S13 is not configured when the EIR only supports Gf)
SGs interface (if Circuit Switched Fallback is required for the MME)
S101 interface (if mobility between eHRPD network is required for the MME)
S102 interface (if CSFB for 1xRTT is required for the MME)
Before the software components are installed, an engineering exercise should determine the number
of compute nodes and VMs for the control plane and data plane. The data plane related VMs should
be configured on the data plane compute nodes whereas the control plane related VMs should be
configured on the control plane compute nodes.
Minimal Steps:
8.2.1 Overview
If the node is currently configured as a standalone MME node, then the following interfaces needs to
be configured to enable the S4-SGSN function:
S3/S16 for SGSN to SGSN/MME mobility management (may include DNS setup)
Additionally, the following interfaces can be optionally configured depending on specific needs:
Gr interface (for roamer from non-EPS capable networks, the UE can only acquire service on the
SGSN side)
Gn interface (if not already configured on the MME and mobility to legacy SGSN or session
management to legacy GGSN is required)
Gf interface (if S13 is not already configured on the MME and the EIR only supports Gf)
When adding SGSN function to an MME, a given set of compute nodes and VMs should be reserved
for SGSN bearer path usage. If there are already MME specific VMs configured on these compute
nodes, they should be removed first. Further, on the remaining compute nodes(the control plane
compute nodes) the subscriber capacity and/or eNodeB fan out capacity may need to be reduced to
leave room to introduce RNC fanout capabilities. To ensure that there is no impact to the subscribers,
it is recommended that the graceful offload function be performed to migrate the subscribers from the
compute nodes to be reconfigured. For example, the following CLI command can be issued to
‘migrate’ the subscribers:
8.3.1 Overview
This section outlines the recommendations for PLMN and pool configuration on the MME/SGSN to
minimize the GUTI/PTMSI collisions. As per 3GPP specification 23.003 Section 2.8.2, when
GUTI/PTMSI conversion are done during inter RAT mobility, the bit 23 to bit 16 of the MTMSI,
corresponding to the NRI byte in PTMSI, are not copied. For example, during GUTI to PTMSI
conversion, this byte within the MTMSI (23rd to 16th) is not copied onto PTMSI, and instead is
replaced with the MME code value from the GUTI. This necessitates a Combo MME/SGSN to have a
uniform allocation schema for GUTI/PTMSI so as to avoid collisions.
The configuration of sgsn-pool is optional for a combined MME/SGSN. If the sgsn-pool is not
configured, all the PLMNs configured on the combo node support all the access types, 2G, 3G, and
4G. When the “default” sgsn-pool is configured, a PLMN must be explicitly configured with sgsn-pool
as “default” to support 2G/3G. If a PLMN is configured with both mme-pool and sgsn-pool, it supports
all the access types (2G/3G/4G). If there is also a PLMN with only sgsn-pool configured as “default”
(mme-pool is “none”), then the PLMN supports only 2G/3G. Since there are two different PLMNs
which is going to use same set of NRI values, there is potential for PTMSI collisions.
Combo/SGSN supports only one SGSN-pool “default” and all the PLMNs configured with this sgsn-
pool uses the same set of NRI values for PTMSI allocation. Therefore, if “default” sgsn-pool is added
to PLMNs having different mme-pool configuration, the PTMSI allocations could potentially collide. In
order to avoid such collisions, the “default” sgsn-pool is recommended to be configured only with one
PLMN or PLMNs with the same mme-pool configuration. Additionally, the mme-codes configured
under the mme-pool should be unique across all the pools.
8.3.2 Example
Example 1
ALERT: PLMNs ‘10067’ and ‘50502’ supports 2G/3G/4G. PLMN ‘50603’ and ‘50631’ are 4G only and
does not support 2G/3G.
Example 2:
ALERT: Both PLMNs having mme-pool 'pool1' supports 2G/3G/4G. PLMNs having 'pool2' does not
support 2G/3G.Removing sgsn-pool from PLMN '10067' converts both the PLMNs in pool to 4G only.
Example 3
ALERT: PLMN ‘10067’ does not support 4G. PLMNs having ‘pool2’ support 4G/3G/2G. However
there is high possibility of GUTI/PTMSI collisions for UEs served by PLMN ‘10067’ and ‘50603’ as
they share same sgsn-pool. Hence this configuration is not recommended.
Example 4
ALERT: PLMN '10067' supports 2G/3G. PLMNs under mme-pool 'pool1' and 'pool2' are 4G only
Example 5
ALERT: If node is COMBO and 'default' sgsn-pool is not configured, then all PLMNs support
2G/3G/4G. There is higher possibility of GUTI/PTMSI collisions for UEs served by PLMN having
pool1 and pool2 and hence this configuration is not recommended. If node is COMBO and ‘default’
sgsn-pool is configured, then all above PLMNs will support only 4G.
To facilitate this, it is recommended to have multiple MME codes configured per MME pool, so that
the combo node uses the same in a round robin fashion.
OAM Mandatory
The following table illustrates the FGW application software functionality to virtual machine mapping.
Step # Action
1 Define the MGMT VMs for RM, PWS and Trace
2 Define the CALLP VMs for the FGW SC, S1ENB and S1MME applications
Each of the 3 applications, is 1+1 spared and runs on each CALLP VM with a spare.
Step # Action
1 Configure MME Pool. Group id, code, relative capacity configured.
user@host% set fgw service-area mme-area 1 name vsim2Area
user@host% set fgw service-area mme-area 1 tac-range-list
0145,1001,1002,51A2,3012,1003,1004,2001,2002,61A2,3014,1004
2 Configure PLMN
Step # Action
1 Configure the SCTP layer.
3 Define S1 Interface (One per FGW CALLP). A loopback interface is required for the S1 interface IP
address.
Step # Action
1 Configure the SCTP layer, and set the rx and tx buffer values to larger than normal SCTP defaults.
2 Define S1 MME Interface. A loopback interface is required for the S1 MME interface IP address.
Step # Action
1 Configure the MME Area.
Step # Action
Step # Action
1 Display the default UE Connection Overload configuration.
Step # Action
1 Display the default FGW Feature Control configuration.
2 Activate Cell Traffic Tracing for HeNB by specifying the HeNB PLMN, HeNB ENB ID, Trace Reference
ID, Trace targets, Trace interfaces and Trace control entity IP.
Trace feature is enabled by default. Operators can active Cell Traffic tracing for 10 HeNBs.
Syntax:
user@host% set fgw trace session <enodeb-plmn-id> enodeb-id <enodeb-id> session-id <session-id>
fgw-s1-interface-mask <fgw-s1-interface-mask> trace-depth <trace-depth> nms-ip <nms-ip>
Step # Action
1 Display the default FGW Health Profile.
Step # Action
Step # Action
1 Unlock the different VMs to bring up the system
2 Verify the status of the VMs and ensure they are all unlocked, the default state for the VM.
The following commands are used to manage the VMs. Note that the <vmtype> is one of ‘mgmt’, 'rm',
‘sig’, 'lb', or ‘data’, and if the VM type is ‘callp’, an extra argument, “<service-id>” is required.
Lock VM (RM/LB/MGMT/DATA/SIG):
→ request mme-sgsn system <vmtype> lock unit-id <VM unit identifier> [force]
Lock VM (CALLP)
request mme-sgsn system callp lock <service-id> unit-id <VM unit identifier>
[force]
Unlock VM (RM/LB/MGMT/DATA/SIG)
Unlock VM (CALLP)
Switchover (CALLP)
Shutdown (MGMT/DATA/SIG)
→ request mme-sgsn system <vmtype> shutdown unit-id <VM unit identifier> [force]
Shutdown (CALLP)
→ request mme-sgsn system callp shutdown <service-id> unit-id <VM unit id> [force]
Warning:
The order in which a partial nodal reset is performed is critical.
4. The standby RM VM
5. The active RM VM
1. The RM VM pair
Per VM information, and relevant high-level OSI administrative and system usage information:
VM storage information:
Enb counts:
Most of these commands are incorporated into a “status” script that can be invoked from the CLI. Use
“describe status” to see the location of the script for customization.
10.4 Backup/Restore
The following procedures describe backup/restore for configuration data of the MME/SGSN VNF. For
software restore, simply reinstall the selected VM images. There are two techniques for doing
configuration backup restore, but in MME8.1, the binary method is the only one that restores the user
accounts completely and LEA config. The XML method is quick for lab settings if user accounts do
not need to be fully restored (i.e. linux layer user setup), and LEA config is not a concern.
10.4.1.1 Backup
user@host> request system recovery backup
Success : Backup request processed. Backup file vmcg40-Apr_23_01:34:47-backup-8.0-23.tar
present at /usr/local/epc/backups
[ok][2015-04-23 01:34:47]
10.4.1.2 Restore
Warning:
A reboot occurs automatically after the restore command is invoked.
10.4.2.1 Backup
user@host> configure
user@host> save <backup-name> xml
ftp from /home/<userid>/<backup-name> to an off-board location
10.4.2.2 Restore
10.5 Security
admin – Can view (show) and change (request/set) all configuration except for lea.
operator – Can view (show) and change (request/set) all configuration except for lea and
aaa/nacm.
lea (law enforcement agency) – Can only view (show) and change (request/set) lea related
configuration.
viewer – Can view all configuration and operational data except for lea and aaa/nacm. This level
can perform show commands, but not request commands.
All the sub-commands associated with Lawful Intercept require a privilege class of lea.
admin@host> configure
Entering configuration mode private
[ok][2014-05-08 23:53:52]
[edit]
admin@host% user ?
Possible completions:
add - Add user account
del - Delete user account
mod - Modify user password
show - Show configured user accounts
admin@host% user add
All changes will be committed at the end of this command. Do you want to continue?
[YES,no]
Enter user name: bob
Enter password:
Confirm password:
Choose a group for the user.
Available groups are: admin lea operator viewer
Enter group for user: operator
Adding user `bob' ...
Adding new user `bob' (1007) with group `operator' ...
Creating home directory `/home/bob' ...
Copying files from `/etc/skel' ...
Adding user 'bob' to the 'operator' group.
User info was successfully added - committing changes.
Commit complete.
[ok][2014-05-08 23:54:36]
[edit]
admin@host% user show
SSH
NAME UID GID PASSWORD KEYDIR HOMEDIR
----------------------------------------------------------------------------------
admin 1001 1001 $1$eVbSG..J$8EKIv9N5l4/OSETuMZNU40 /home/admin
bob 1007 1003 $1$7avF083L$Q0/xOjaEdmH2hZqN2z6GT. /home/bob
lea 1002 1002 $1$3b0jvqhP$TgsMGFgQOA38DX1kNlO170 /home/lea
operator 1003 1003 $1$603BmPyE$ij7cItjw5rScGAa519IXt/ /home/operator
viewer 1004 1004 $1$wY9m84vW$l1VEE9qxWTAZ6QgJPx/FA0 /home/viewer
[ok][2014-05-08 23:54:53]
[edit]
admin@host%
admin@host> configure
Entering configuration mode private
[ok][2014-05-09 06:05:06]
[edit]
admin@host% user del
All changes will be committed at the end of this command. Do you want to continue?
[YES,no]
Enter user name: bob
gid 1003
Deleting user
Deleting user from the operator group.
Looking for files to backup/remove ...
Removing files ...
Removing user `bob' ...
Warning: group `operator' has no more members.
Done.
User info was successfully deleted - committing changes.
Commit complete.
[ok][2014-05-09 06:05:11]
[edit]
admin@host%
[ok][2014-05-09 06:05:52]
[edit]
admin@host%
Note: only admin group users have permission to see the contents of this file