Cisco GGSN Release 8.0 Configuration Guide, Cisco IOS Release 12.4(15)XQ

Configuring Charging on the GGSN

This chapter describes how to configure the charging function on a gateway GPRS support node (GGSN). If at minimum, one charging gateway is configured, by default, charging processing is enabled on the GGSN. There are several ways to customize communication with a charging gateway. Many of the default values for the charging options will provide a satisfactory configuration until you become more familiar with your network and decide to customize the charging interface.

For a complete description of the GGSN commands in this chapter, refer to the Cisco GGSN Command Reference for the Cisco GGSN release you are using. To locate documentation of other commands that appear in this chapter, use the command reference master index or search online.

This chapter includes the following sections:

•Configuring an Interface to the Charging Gateway (Required)

•Configuring the Default Charging Gateway (Required)

•Configuring a Charging Source Interface (Optional)

•Configuring the GGSN Memory Threshold (Optional)

•Configuring the Transport Protocol for the Charging Gateway (Optional)

•Configuring the Charging Release (Optional)

•Configuring Charging for Roamers (Optional)

•Customizing the Charging Gateway (Optional)

•Disabling Charging Processing (Optional)

•Using Charging Profiles (Optional)

•Configuring G-CDR Backup and Retrieval using iSCSI (Optional)

•Monitoring and Maintaining Charging on the GGSN

•Configuration Examples

Configuring an Interface to the Charging Gateway

To establish access to an external charging gateway in the general packet radio service/Universal Mobile Telecommunication System (GPRS/UMTS) network, you must configure a interface on the GGSN to connect to the network of the charging gateway. In GPRS/UMTS, the interface between the GGSN and the charging gateway is referred to as the Ga interface. The Cisco GGSN supports both a 2.5G Ga interface and 3G Ga interface.

On the Cisco 7600 series router platform, this interface is logical one (on which IEEE 802.1Q-encapsulation has been configured) to the Layer 3 routed Ga VLAN configured on the supervisor engine.

For more information about the Ga VLAN on the supervisor engine, see "Platform Prerequisites" section.

For more information about configuring interfaces, see the Cisco IOS Interface Configuration Guide and the Cisco IOS Interface Command Reference.

Configuring 802.1Q-Encapsulated Subinterfaces

To configure a subinterface that supports IEEE 802.1Q encapsulation to the Ga VLAN, use the following commands, beginning in global configuration mode:

Verifying Interface Configuration to the Charging Gateway

To verify the interface to the charging gateway (CG) you can first verify your GGSN configuration and then verify that the interface is available.

Step 1 To verify that you have properly configured a Ga interface on the supervisor engine, use the show running-config command. The following example is a portion of the output from the command showing the Fast Ethernet 8/22 physical interface configuration as the Ga interface to the SGSN. The configuration of the Fast Ethernet 8/22 physical interface is shown in bold.

Sup# show running-config

Building configuration...

Current configuration :12672 bytes

!

version 12.2

...

interface FastEthernet8/22

 no ip address

 switchport

 switchport access vlan 302

!

interface Vlan101

 description Vlan to GGSN for GA/GN

 ip address 10.1.1.1 255.255.255.0

!

interface Vlan302

 ip address 40.0.2.1 255.255.255.0

Step 2 To verify that the physical interface and the Ga VLAN are available, use the show interface command on the supervisor engine. The following example shows that the Fast Ethernet 8/22 physical interface to the charging gateway is up as well as the Ga VLAN, VLAN 101:

Sup# show ip interface brief FastEthernet8/22

Interface                  IP-Address      OK? Method Status                Protocol

FastEthernet8/22           unassigned      YES unset  up                    up 

Sup# show ip interface brief Vlan302

Interface                  IP-Address      OK? Method Status                Protocol

Vlan302                    40.0.2.1        YES TFTP   up                    up      

Sup#

Step 3 To verify the Ga VLAN configuration and availability, use the show vlan name command on the supervisor engine. The following example shows the Gn VLAN Gn_1:

Sup# show vlan name Ga_1

VLAN Name                             Status    Ports

---- -------------------------------- --------- -------------------------------

302  Ga_1                             active    Gi4/1, Gi4/2, Gi4/3, Gi7/1

                                                Gi7/2, Gi7/3, Fa8/22, Fa8/26

VLAN Type  SAID       MTU   Parent RingNo BridgeNo Stp  BrdgMode Trans1 Trans2

---- ----- ---------- ----- ------ ------ -------- ---- -------- ------ ------

302  enet  100302     1500  -      -      -        -    -        0      0   

Remote SPAN VLAN

----------------

Disabled

Primary Secondary Type              Ports

------- --------- ----------------- ------------------------------------------ 

Step 4 On the GGSN, to verify that you have properly configured a Ga subinterface to the Ga VLAN, use the show running-config command. The following example is a portion of the output from the command which shows a Fast Ethernet 5/1 physical interface configuration as the Ga interface to the charging gateway:

GGSN# show running-config

Building configuration...

Current configuration :7390 bytes

!

! Last configuration change at 16:56:05 UTC Wed Jun 25 2003

! NVRAM config last updated at 23:40:27 UTC Fri Jun 13 2003

!

version 12.3

.....

interface GigabitEthernet0/0.2

 description Ga/Gn Interface

 encapsulation dot1Q 101

 ip address 10.1.1.72 255.255.255.0

 no cdp enable

!

.....

ip route 40.1.2.1 255.255.255.255 10.1.1.1

Step 5 To verify that the subinterface is available, use the show ip interface brief command. The following example shows that the Gigabit Ethernet 0/0.2 subinterface to the Ga VLAN is in "up" status and the protocol is also "up":

GGSN# show ip interface brief GigabitEthernet0/0.2

Interface                  IP-Address      OK? Method Status                Protocol

GigabitEthernet0/0.2       10.1.1.72       YES NVRAM  up                    up 

Configuring the Default Charging Gateway

You can configure a primary charging gateway that the GGSN uses, by default, to communicate charging information. Additionally, you can specify a secondary and tertiary charging gateway as backups. All charging gateways share the same global charging parameters.

To configure a default charging gateway for a GGSN, use the following command in global configuration mode:

Configuring the GGSN to Switchover to the Highest Priority Charging Gateway

When priority switchover has been configured on the GGSN using the gprs charging switchover priority command, regardless of the state of the current active charging gateway, when a gateway of higher priority comes up, the GGSN will switch over and send G-CDRs to that charging gateway.

To configuring priority switchover on the GGSN, use the following command in global configuration mode:

Changing the Default Charging Gateway

To change the default charging gateway of a GGSN, use the following commands, beginning in global configuration mode:

Configuring a Charging Source Interface

By default, the global GTP virtual template interface is used for all charging messages. With Cisco GGSN Release 8.0 and later, you can configure a charging source interface to use for charging messages.

The charging source interface is a loopback interface that the GGSN has been configured to use for charging traffic using the gprs charging interface source loopback global configuration command. Once a loopback interface has been configured as the charging source interface, all charging messages will use the IP address of that loopback interface as their source address.

This feature enables you to separate charging traffic. Optionally, VRF can be configured on the loopback interface, which enables charging traffic to be separated onto a private VLAN.

When configuring a charging source interface, note the following:

•Once configured, the loopback interface cannot not be modified without removing the charging source interface configuration. All charging messages will use the new end points from the path structure.

•A charging source interface cannot be unconfigured while there are active PDPs or CDRs.

To configure a charging source interface, complete the following tasks, beginning in global configuration mode:

Router(cfg-acct-mlist)# exit

To configure the GGSN to use the loopback interface for charging traffic, use the following command while in global configuration mode:

Router(config)# gprs charging interface source 
loopback number

Configuring the GGSN Memory Threshold

The GGSN memory protection feature prevents processor memory from being drained during periods of abnormal conditions (such as when all charging gateways are down and the GGSN is buffering CDRs into memory. By default, the memory threshold is 10% of the total memory available at the time GGSN services are enabled using the gprs ggsn service global configuration command. You can use the gprs memory threshold global configuration command to configure the threshold according to the router and memory size.

When the amount of memory remaining on the system reaches the defined threshold, the memory protection feature activates and the GGSN performs the following actions to keep the processor memory from falling below the threshold:

•Rejects new create PDP requests with the cause value "No Resource."

•Drops any existing PDPs for which an update is received with the cause value "Management Intervention."

•Drops any PDPs for which a volume trigger has occurred.

Note While the memory protection feature is active, byte counts will be maintained and reported after the GGSN recovers. However, because some change conditions are not handled, some counts will not reflect the accurate charging condition (for example, QoS and tariff conditions).

To configure the memory threshold that when reached, activates the memory protection feature on the GGSN, use the following global configuration command:

Configuring the Transport Protocol for the Charging Gateway

You can configure a GGSN to support either Transport Control Protocol (TCP) or User Datagram Protocol (UDP) as the transport path protocol for communication with the charging gateway.

The GGSN default configuration specifies UDP, which is a connectionless protocol that is considered an unreliable transport method but can yield greater performance.

Configuring TCP as the Charging Gateway Path Protocol

TCP is a connection-based protocol that provides reliable transmission through packet acknowledgment. To specify TCP as the transport path protocol, use the following commands, beginning in global configuration mode:

Configuring UDP as the Charging Gateway Path Protocol

The GGSN default configuration specifies UDP as the transport path protocol to the charging gateway. If you need to reconfigure the charging gateway for UDP transport, use the following command in global configuration mode:

Configuring the Charging Release

The Cisco GGSN supports both 2.5G and 3G Ga interfaces and GPRS (R97/R98) and UMTS (R99) Quality of Service (QoS) profile formats. The Cisco GGSN can be configured to comply with 3GPP TS 32.215 Release 4 or Release 5.

Depending on the CG and GGSN configuration, when specifying the 99 or 98 keyword, the following actions take place:

•If the GGSN is configured to present R97/R98 CDRs (gprs charging release 98 is configure):

–If the PDP context is R98, the GGSN presents an R97/R98 G-CDR.

–If the PDP context is R99, the GGSN presents an R97/R98 G-CDR by converting the R99 QoS profile to an R97/R98 QoS profile.

•If the GGSN is configured to present R99 CDRs (gprs charging release 99 is configure):

–If the PDP context is R99, the GGSN presents an R99 G-CDR.

–If the PDP context is R98, the GGSN presents an R99 CDR by converting the QoS profile.

To configure the charging release with which the GGSN complies when presenting G-CDRs, use the following command in global configuration mode:

Configuring Charging for Roamers

A GGSN can be configured to generate G-CDRs for roaming mobile subscribers.

When the charging for roamers feature is enabled on the GGSN, when the GGSN receives a PDP context request, it first checks to see if both the GGSN and serving GPRS support node (SGSN) public land mobile network (PLMN) IDs are present and match (via the Routing Area Identity [RAI] field information element [IE]).

If not both are not present and match, the GGSN matches the IE containing the SGSN Signaling Address field against a list of PLMN IP address ranges that have been defined using the gprs plmn ip address command with the sgsn keyword option specified.

Note To use the RAI IE in Create PDP Context requests to detect roamers, a valid home PLMN must be configured on the GGSN using the gprs mcc mn global configuration command. When a valid home PLMN is configured, or valid trusted PLMNs, a CDR will not be generated if the RAI matches the configured home (or trusted) PLMN. A CDR will be created for all PDPs with RAIs that do not match a home or trusted PLMN.

Note If the RAI field is not present in a Create PDP Context, and an address range has not been configured using the gprs plmn ip address command with the sgsn keyword option specified, the PDP will be classified as "unknown" and treated as a roamer.

If the GGSN determines that the SGSN that sent the Create PDP Context request is not located within the same PLMN as it is, the GGSN generates a call detail record (CDR). If the GGSN determines that the SGSN is located in the same PLMN, it will not generate a CDR until it receives notification that the SGSN has changed location to another PLMN.

To enable charging for roamers on the GGSN using the gprs charging roamers command, you should first define a set of IP address ranges for a PLMN, using the gprs plmn ip address command.

Note It is important that you configure the gprs plmn ip address and gprs charging roamers commands in their proper order. After you configure the IP address range for a PLMN, use the gprs charging roamers command to enable the charging for roamers feature on the GGSN. You can change the IP address range by reissuing the gprs plmn ip address command.

To verify your configuration, use the show gprs charging parameters command to see if the charging for roamers feature is enabled. To verify your PLMN IP address ranges, use the show gprs plmn ip address command.

Configuring PLMN IP Address Ranges

Depending on how the PLMN IP address ranges have been defined using the gprs plmn ip address start_ip end_ip [sgsn] command, the charging for roamers feature operates as follows:

•If no PLMN IP address ranges are configured using the gprs plmn ip address start_ip end_ip [sgsn] command, the GGSN generates CDRs for all initiated PDP contexts regardless of whether the GGSN and SGSN are located within the same PLMN.

•If a list of PLMN IP address ranges has been configured using the gprs plmn ip address start_ip end_ip [sgsn] command, and one or more of those ranges has been defined using the sgsn key word, the GGSN uses those ranges defined with the sgsn keyword to determine whether an SGSN is located within the same PLMN.

With this configuration, the following scenarios outline how the charging for roamers feature will function:

–MS1 is subscribed to PLMN1 and attaches to an SGSN in PLMN2. From PLMN2, MS1 initiates a PDP context with the GGSN in PLMN1. In this case, MS1 is a roamer and the GGSN generates a CDR because it determines that the SGSN is located in a different PLMN.

–MS1 is subscribed to PLMN1 and attaches to an SGSN in PLMN2. From PLMN2, MS1 initiates a PDP context with the GGSN in PLMN2. In this case, MS1 is not a roamer because the SGSN and GGSN are in the same PLMN. The GGSN does not create a G-CDR.

To configure PLMN IP address ranges, use the following command in global configuration mode:

Enabling Charging for Roamers

To enable the charging for roamers feature on a GGSN, use the following command in global configuration mode:

Customizing the Charging Gateway

For the GGSN charging options, the default values represent recommended values. Other optional commands are also set to default values; however, we recommend modifying these commands to optimize your network as necessary, or according to your hardware.

The GGSN uses echo timing to maintain the path between SGSNs and external charging gateways. However, the GGSN can implement only a single method of echo timing for all the paths that it needs to maintain. To learn more about echo timing on the GGSN, or to modify the echo timing feature, see the "Configuring Echo Timing on a GGSN" section in the "Configuring GTP Services on the GGSN" chapter.

Use the following global configuration commands to fine-tune charging processing on the GGSN:

Router(config)# gprs charging 
cdr-aggregation-limit CDR_limit

Router(config)# gprs charging cdr-option apn 
[virtual]

Router(config)# gprs charging cdr-option 
apn-selection-mode 

Router(config)# gprs charging cdr-option 
camel-charge-info

Router(config)# gprs charging cdr-option 
chch-selection-mode

Router(config)# gprs charging cdr-option 
dynamic-address

Router(config)# gprs charging cdr-option imeisv

Router(config)# gprs charging cdr-option 
local-record-sequence-number 

Router(config)# gprs charging cdr-option 
ms-time-zone

Router(config)# gprs charging cdr-option nip

Router(config)# gprs charging cdr-option 
no-partial-cdr-generation [all]

Router(config)# gprs charging cdr-option 
node-id

Router(config)# gprs charging cdr-option 
packet-count

Router(config)# gprs charging cdr-option 
pdp-address

Router(config)# gprs charging cdr-option 
pdp-type

Router(config)# gprs charging cdr-option 
rat-type

Router(config)# gprs charging cdr-option 
served-msisdn

Router(config)# gprs charging cdr-option 
service-record [value]

Router(config)# gprs charging cdr-option 
sgsn-plmn

Router(config)# gprs charging cdr-option 
user-loc-info

Router(config)# gprs charging 
cg-path-requests minutes

Router(config)# gprs charging container 
volume-threshold threshold_value

Router(config)# gprs charging disable

Router(config)# gprs charging flow-control 
private-echo

Router(config)# gprs charging header short

Router(config)# gprs charging map data tos 
tos_value

Router(config)# gprs charging message 
transfer-request possibly-duplicate

Router(config)# gprs charging packet-queue-size 
queue_size

Router(config)# gprs charging path-protocol 
{udp | tcp}

Router(config)# gprs charging port port-num

Router(config)# gprs charging send-buffer bytes

Router(config)# gprs charging 
server-switch-timer seconds

Router(config)# gprs charging tariff-time time

Router(config)# gprs charging message 
transfer-request command-ie

Router(config)# gprs charging message 
transfer-response number-responded

Router(config)# gprs charging reconnect minutes

Router(config)# gprs charging transfer interval 
seconds

For information about configuring GGSN GTP options, see the "Customizing the GGSN Configuration" section in the "Configuring GTP Services on the GGSN" chapter.

Disabling Charging Processing

Caution The gprs charging disable command removes charging data processing on a GGSN, which means that the data required to bill customers for network usage is neither being collected by the GGSN nor being sent to the charging gateway. We recommend that you avoid using this command in production GPRS/UMTS network environments. When it is necessary to use this command, use it with extreme care and reserve its usage only under nonproduction network conditions.

You can disable charging on the GGSN only after all the open CDRs have been processed and sent to the charging gateway. To clear the current GGSN CDRs, use the clear gprs charging cdr privileged EXEC command.

To disable charging processing on a GGSN, use the following command, beginning in global configuration mode:

Using Charging Profiles

You can apply different charging methods on a per-PDP basis using charging profiles that you create, customize, and specify as the default charging method to use for a specific type of user at an APN level and global level. Charging profiles provide the ability to offer flexible services that are customized to subscriber preferences.

When using charging profiles, note the following:

•The GGSN must be configured to include the charging characteristics selection mode parameter in CDRs using the gprs charging cdr-option chch-selection-mode global configuration command.

•The GGSN must be configured to receive the charging characteristics selection mode IE in CDRs by using the gprs charging release global configuration command.

To apply charging methods on a per-PDP basis using GGSN charging profiles, you must complete the tasks outline in the following sections:

•Configuring a Charging Profile

•Defining the Charging Characteristics and Triggers of the Charging Profile

•Applying a Default Charging Profile to an APN

•Applying a Global Default Charging Profile

•Configuring How the GGSN Handles PDPs with Unmatched Charging Profiles

Configuring a Charging Profile

Charging profiles define the charging method to apply to a specific type of user (home, roamer, visitor).

The GGSN supports up to 256 charging profiles numbered 0 to 255.

Profile 0 is a set profile that always exists on the GGSN. It is not created by a GGSN operator, however, it can be modified using the charging-related global configuration commands. Profiles 1 to 255 are user-defined and customized using charging profile configuration commands.

When a Create PDP Context request is received, an appropriate charging profile is selected based on the following sources of input:

•SGSN/HLR via the charging characteristics IE.

•Local defaults.

•Charging profile index AAA attribute.

Note The charging profile index received from AAA will take effect only if service-awareness has been configured globally on the GGSN (using the gprs service-aware global configuration command), and at the APN level (using the service-aware access-point configuration command).

For information on configuring a service-aware GGSN, see the "Configuring Enhanced Service-Aware Billing" chapter of the Cisco GGSN Configuration Guide.

The order in which a charging profile is selected for a PDP context, is as follows:

1. Charging profile index in the override rule on the APN—If a default charging profile has been configured at both the APN and global level to override the SGSN specification, the APN default charging profile is used first.

2. Charging profile index in the override rule on the box (global default charging profile)—If there is no default charging profile default configured at the APN, the default charging profile configured globally is use.

3. Charging profile index from AAA.

4. Charging profile index from SGSN/HLR

5. Charging profile index from the non-override rule on the APN.

6. Charging profile index from non-override rule on the box (global default charging profile).

If none of the above applies, the PDP context is rejected if the gprs charging characteristics reject global configuration command is configured and the create request is GTP v1. If the gprs charging characteristics reject command is not configured, the GTPv1 PDP context is created using charging profile 0.

Note The default charging profile, i.e. charging profile 0, is not supported for service-aware PDPs. These PDP create requests will be rejected with error code 199.

Note You cannot remove a charging profile, or modify a DCCA profile under a charging profile (using the content dcca profile charging profile command) if there are any PDPs contexts or rules using the profile. If rules are using the profile, you must disassociate the profile with the APN before making configuration changes.

To create or modify a charging profile and enter charging profile configuration mode, use the following global configuration command:

Defining the Charging Characteristics and Triggers of the Charging Profile

To configure the charging methods and triggers of a charging profile, use the following commands in charging profile configuration mode:

Applying a Default Charging Profile to an APN

To configure a default charging profile to use for a specific type of user at an APN, use the following access-point configuration command:

Applying a Global Default Charging Profile

Default charging profiles configured at the global level are used when a default charging profile has not been specified for an APN.

To configure a default charging profile to use for a specific type of user globally, use the following global configuration command:

Configuring How the GGSN Handles PDPs with Unmatched Charging Profiles

The GGSN can be configured to reject or accept GTPv1 Create PDP Context requests for which a profile cannot be matched. If configured to accept these PDP context requests, the charging method defined by charging profile 0 is applied. By default, the Create PDP Context requests are accepted and the charging method defined in charging profile 0 is applied.

The following restrictions apply to charging profiles selected for service-aware PDPs:

•All PDP s belonging to the same user must use the same charging profile as that of the primary PDP.

•The default charging profile, i.e. charging profile 0, is not supported for service-aware PDPs. These PDP create requests will be rejected with error code 199.

To configure a GGSN to reject Create PDP Context requests for which a charging profile cannot be matched, use the following global configuration command:

Configuring G-CDR Backup and Retrieval using iSCSI

The Cisco GGSN Release 8.0 and later utilizes the Cisco IOS software Small Computer Systems Interface over IP (iSCSI) support, as defined in RFC 3720, to enable G-CDR storage and retrieval from storage target on a Storage Area Network (SAN).

This section contains the following topics on iSCSI support on the GGSN:

•iSCSI Overview

•Configuring iSCSI Backup and Storage on the GGSN

•Monitoring and Maintaining iSCSI G-CDR Backup and Storage

iSCSI Overview

The iSCSI transport protocol operates over TCP/IP, enabling mobile operators and service providers to use their SAN connected to an iSCSI interface to save complete DTR messages containing closed CDRs.

SAN technology, which enables customers to build scalable storage solutions, is comprised of the following primary elements:

•SCSI—An interface standard which enables multiple devices to be installed on a system, attached to cable to form a chain of devices. Each device is assigned a unique ID, which is expressed as a number, that identifies that device on the bus. SCSI IDs can be broken into Logical Unit Numbers (LUNs), enabling a number of devices to share a single SCSI ID. Devices from which I/O requests originate are called initiators, and devices from which responses originate are called targets.

•SAN—Technology that involves moving network storage to a separate network of its own. Disk, tape, and optical storage can then be attached to the storage network that is based on a fabric of switches and hubs that connects storage devices to a heterogeneous set of servers.

A SAN system provides block-level access to data residing on shared storage arrays through dedicated storage networks.

•iSCSI—Transport protocol that maps SCSI requests and responses over TCP and provides block-level data transfer between the SCSI initiator (the Cisco GGSN is this example), and the target (the storage device on the SAN). The initiator sends I/O requests and the target sends I/O responses.

A SAN topology is distinguished by the following features

•Storage is not directly connected to network clients

•Storage is not directly connected to servers

•Storage devices are interconnected

•Multiple servers can share multiple storage devices

Configuring iSCSI Backup and Storage on the GGSN

In the SCSI environment, the GGSN functions as an iSCSI initiator.

To enable G-CDR backup storage on an iSCSI device, you must first configure an iSCSI target profile on the GGSN that includes the name and IP address of the target, and the TCP port on which to "listen" for iSCSI traffic. Once the profile has been configured, you then configure the GGSN to use the interface for record storage when a charging gateway is not available.

With Cisco GGSN Release 9.0 and later, up to 30 iSCSI target profiles can be configured and associated with a set of unique charging gateways within a charging group.

As an alternative, an iSCSI target profile can be configured as the primary storage for CDRs by only configuring an iSCSI target profile and no charging gateway at the global level (default charging group 0), or at the APN level by defining only an iSCSI target in the charging group associated with an APN (charging groups 1 to 29).

The I/O requests sent by the GGSN are converted into SCSI requests and transported over TCP/IP to the remote storage target.

Choosing the Record Format when Writing to iSCSI

By default, when writing DTRs to iSCSI, the format for storing records is "GTP," wherein the complete DTR is written to the iSCSI target. As an alternative, the iSCSI record format can be configured as ASN.1 using the gprs charging iscsi rec-format command in global configuration mode and specifying the asn.1 keyword option. When ASN.1 is configured as the record format, the GGSN writes only the raw ASN1-encoded CDRs into iSCSI without embedding the DTR information element into the records. This format is useful when the records are retrieved from iSCSI using FTP.

To configure the record format, use the gprs charging iscsi rec-format command in global configuration mode.

Writing DTRs when iSCSI is used as Backup Storage

•Once iSCSI backup storage configuration is in place, when a charging gateway is not reachable, the writing towards the iSCSI is initiated. The complete DTR message is sent to the iSCSI target defined in the target profile.

•The recommended iSCSI record format when using iSCSI as backup is GTP (the default format). If the iSCSI auto-retrieval (gprs auto-retrieve global configuration command) is enabled, the record format must be configured as GTP.

•When iSCSI auto-retrieval is enabled, along with sending the complete DTR message, the GGSN adds a 12-byte header in front of the message before storing it to SAN. This header is used when the DTRs are retrieved and sent to the charging gateway. (In addition, the RSM-layer adds a 12-byte header and 4-byte trailer to the message before it is stored).

Note If the DTRs are retrieved directly from the SAN through other means such as FTP, then each record must skip the 10-byte header to get to the actual DTR containing encoded CDRs.

Writing DTRs when iSCSI is used as Primary Storage

•When there are no charging gateways configured, and only an iSCSI target profile is defined at the global charging level (charging group 0), or granular charging level (charging groups 1 to 29), the iSCSI is the primary storage for writing charging records.

•Although any iSCSI record format can be used, the ASN.1 iSCSI record format would enable the user to store raw ASN.1 encoded CDRs into iSCSI without any additional headers.

Note The records in ASN.1 format are generated only when gprs auto-retrieve is disabled on the GGSN, which is the default behavior. The ASN.1 format should be used only when the iSCSI target is used as the primary storage for charging records (no charging gateways are configured).

Reading CDRs

•Once the iSCSI backup storage configuration is in place, when a charging gateway comes up, the iSCSI initiator (GGSN) will request for any iSCSI records to be received from the iSCSI target.

•Once the GGSN receives a record, the 12-byte header added by the GGSN when the write process is removed and the complete DTR is sent to the charging gateway.

If the DTRs are to be marked for possibly duplicate before sending to the charging gateway, GGSN has to be configured with the following charging config command.

–gprs charging message transfer-request command-ie

–gprs charging message transfer-request possibly-duplicate

iSCSI Restrictions

When configuring iSCSI G-CDR backup and storage on the GGSN, note the following:

•Currently, iSCSI targets cannot be dynamically discovered.

•The number of TCP connections per iSCSI session is limited to one.

•The iSCSI target device should be preformatted. Each LUN must have only one FAT32 partition.

•Maximum of size of a LUN must not be more than 2TB, which is the maximum disk size supported by a FAT32 file system.

When configuring iSCSI G-CDR Backup and Storage on the GGSN, complete the tasks in the following sections:

•Configuring an iSCSI Target Interface Profile

•Associating an iSCSI Target Interface Profile

•Verifying the iSCSI Session

Configuring an iSCSI Target Interface Profile

Note Up to 30 iSCSI profiles can be configured on the GGSN, however, only one target can be defined per profile, and only one profile at a time can be associated with the GGSN to use the iSCSI interface using the gprs iscsi command in global configuration mode.

To configure the iSCSI target interface profile on the GGSN, complete the following tasks, beginning in global configuration mode:

Router(config)# ip iscsi target-profile 
target_profile_name

Router(config-iscsi)# name target name

Router(config-iscsi)# ip ip_address

Router(config-iscsi)# port tcp port

Router(config-iscsi)# source-interface 
loopback_interface_number

Router(config-iscsi)# vrf vrf_name

Router(config-iscsi)# exit

Note The name, ip, and port iSCSI interface sub configurations are required. For a complete list of optional configurations that you can configure under a target interface profile, issue the "?" command while in iSCSI interface configuration mode or see the ip iscsi target-profile command description in the Cisco GGSN Release 8.0 Command Reference.

Associating an iSCSI Target Interface Profile

To configure the GGSN to use the iSCSI interface for G-CDR storage when no charging gateway is available, use the following command in global configuration mode:

Router(config)# gprs iscsi target_profile_name

Verifying the iSCSI Session

To verify that the iSCSI session is up, use the following command in privileged EXEC mode:

Router# show ip iscsi session

Monitoring and Maintaining iSCSI G-CDR Backup and Storage

The following is a list of commands that you can use to monitor and maintain the iSCSI backup and storage functions on the GGSN:

Router# show record-storage-module target-info [all | 
target-profile profile_name]

Monitoring and Maintaining Charging on the GGSN

This section provides a summary list of the show commands that you can use to monitor charging functions on the GGSN.

The following privileged EXEC commands are used to monitor and maintain charging on the GGSN:

Configuration Examples

The following are examples of charging configurations implemented on the GGSN.

Global Charging Configuration

GGSN Configuration

Router# show running-config

Building configuration...

Current configuration :7390 bytes

!

! Last configuration change at 16:56:05 UTC Wed Jun 25 2003

! NVRAM config last updated at 23:40:27 UTC Fri Jun 13 2003

!

version 12.3

.....

interface GigabitEthernet0/0.2

 description Ga/Gn Interface

 encapsulation dot1Q 101

 ip address 10.1.1.72 255.255.255.0

 no cdp enable

!

.....

ip route 40.1.2.1 255.255.255.255 10.1.1.1

!

gprs access-point-list gprs

  access-point 1

   access-point-name auth-accounting

   access-mode non-transparent

   aaa-group authentication first

   aaa-group accounting second

   ip-address-pool dhcp-proxy-client

   dhcp-server 10.60.0.1

   dhcp-gateway-address 10.60.0.1       

   exit

   !

. . .

!

gprs default charging-gateway 10.9.0.2

gprs charging send-buffer 1000

gprs charging container volume-threshold 500000

gprs charging container change-limit 3

gprs charging cdr-aggregation-limit 10

gprs charging cdr-option apn-selection-mode

gprs charging cdr-option served-msisdn

!

gprs memory threshold 512

!

. . .

!

end

Supervisor Engine Configuration

Sup# show running-config

Building configuration...

Current configuration :12672 bytes

!

version 12.2

...

interface FastEthernet8/22

 no ip address

 switchport

 switchport access vlan 302

!

interface Vlan101

 description Vlan to GGSN for GA/GN

 ip address 10.1.1.1 255.255.255.0

!

interface Vlan302

 ip address 40.0.2.1 255.255.255.0

Charging Profile Configuration

The following partial configuration example shows two charging profiles (charging profile 1 and charging profile 2) configured on the GGSN, with charging profile 1 being configured as the global default charging profile to be used for "any" type of user if a charging profile is not specified at the APN:

Router# show running-config

Building configuration...

Current configuration :7390 bytes

!

! Last configuration change at 16:56:05 UTC Wed Jun 25 2003

! NVRAM config last updated at 23:40:27 UTC Fri Jun 13 2003

!

version 12.3

.....

interface GigabitEthernet0/0.2

 description Ga/Gn Interface

 encapsulation dot1Q 101

 ip address 10.1.1.72 255.255.255.0

 no cdp enable

!

.....

ip route 40.1.2.1 255.255.255.255 10.1.1.1

!

!

. . .

!

gprs charging profile default any 1

gprs charging profile 1

 description "roamer_profile"

 limit volume 500000 reset

 limit duration 30 reset

!

gprs charging profile 2

 description "any_unmatched"

 limit volume 1000000 reset

 limit duration 60 reset

. . .

!

. . .

!

end