Digital Trunks

Digital trunks typically offer much greater capacity, quality, and reliability than analog trunks. As such, most midsize and larger businesses use digital service connections to the PSTN. This rise in popularity has also caused the cost of this service to decline over time in most geographies, making its use even more attractive.

In this section, you will learn how to configure, verify, and troubleshoot the four types of digital service that Cisco voice gateways support. These include ISDN PRI, ISDN BRI, E1 R2, and T1 CAS trunk types.

Configuring E1/T1 Physical Layer Connections

All these trunk typeswith the exception of ISDN BRIare delivered on either an E1 or T1 digital circuit. Before you configure voice signaling methods, be sure that the interface is configured so that the circuit is operating properly on the physical signaling level (Layer 1).

E1 or T1 voice circuits are channelized, meaning they transmit information in timeslots or channels. E1 circuits have 30 timeslots available, and T1 circuits have 24 timeslots. The timeslots are carried in frames. The beginning of the frame is marked by a specific bit sequence. A more complete description of the circuit characteristics is described in Chapter 5, "Circuit Options."

At a physical level, the interface must be able to identify the beginning of each frame in the bit stream and understand the frame format that is being used so that it can read the information in the timeslots correctly. This means several things:

• The interface must be accurately clocked to prevent bit "slips." A slip occurs when there is a difference between when the interface expects a data bit and when it actually arrives. When connecting to the PSTN, you should always use line clockingthat is, obtain the clock from the PSTN circuit. PSTN equipment uses extremely accurate clock sources and provides an excellent source for precisely clocking the voice gateway interface.
• T1 or E1 framing is necessary to provide a common data format and to provide a means for synchronization on a network. For T1, two common framing standards are currently in use: Super Frame (D4) and Extended Superframe (ESF). An E1 circuit uses timeslot 0 for framing and synchronization. Timeslot 0 has two framing formats: double-frame and CRC-4. Double-frame provides synchronization but no error checking. CRC-4 adds a checksum to provide error checking of timeslots within the frame.
• Line coding identifies the way the bit stream is sent on the transmission line. T1 circuits can use either alternate mark inversion (AMI) or binary 8-zero substitution (B8ZS) line coding. E1 circuits can also use AMI, or they can use high-density Bipolar 3 (HDB3) line coding.

For T1, D4 framing is usually associated with AMI line coding, and ESF framing is usually associated with B8ZS line coding. The PSTN service provider can furnish details regarding which framing and line coding mechanism is being used on the circuit that you are installing.

Now you can begin setting up the physical characteristics of the circuit at the T1/E1 controller level. From global configuration mode, enter the controller T1 port/slot or controller E1 port/slot command to enter controller configuration mode.

Configure the framing using the framing type command. For T1, the valid options for type are sf or esf. The default for T1 is SF. For E1, the valid options for type are crc4 and no-crc4. The default for E1 is CRC4.

Configure the line coding using the linecode type command. For T1, the valid options for type are ami and b8zs. The default for T1 is AMI. For E1, the valid options for type are ami and hdb3. The default for E1 is HDB3.

It is not normally necessary to configure the clock source for PSTN connections. The default setting is line clock, which allows the interface to obtain clocking from the circuit. This is typically the correct setting for a PSTN circuit.

Example 6-10 demonstrates configuring the physical interface for an E1 circuit.

After you have configured the proper framing and line coding, the controller comes up.

Verifying and Troubleshooting

You can verify that no physical alarms exist on this circuit by using the show controller T1 slot/port or show controller E1 slot/port command. Both commands show you the current state of the controller, configuration information, and whether any alarms are detected. Both commands also provide error counters for the previous 24 hours in 15-minute intervals.

Example 6-11 demonstrates sample output from the show controller e1 command. You can see in this example that this interface has had errors. Slip seconds have been recorded in the current interval, indicating a possible clocking issue.

Table 6-1 describes all the fields shown.

Configuring ISDN PRI Trunks

To configure a PRI trunk, you need to gather the following information from the PSTN service provider:

• Number of channels This is the number of channels provisioned for voice service.
• ISDN channel selection order Determine if the PSTN service provider expects B-channel selection to be ascending (B-channel 1 first) or descending (B-channel 30 first for E1 or B-channel 23 first for T1). The default selection order on the Cisco voice gateways is descending.
• ISDN switch type Specifying an ISDN switch type identifies the Layer 2 and Layer 3 signaling that the voice gateway and the PSTN are using. The ISDN switch types that Cisco voice gateways support vary depending on the Cisco IOS software release. Table 6-2 lists the switch types that Cisco IOS Software Release 12.3 supports.

Basic configuration of the interface for a PRI voice trunk begins by configuring the gateway for an ISDN switch type. You do this in global configuration mode by using the isdn switch-type switch type command, where switch type matches the device that the PSTN service provider is using; see Table 6-2. Defining a switch type globally for the gateway is required.

Note

You can override the ISDN switch type at the individual interface level. This is useful when you have interfaces in the gateway that are connected to different switches.

The second step is to configure the PRI voice timeslots on the controller. Enter controller configuration and use the command pri-group timeslots (1-23) for T1 or pri-group timeslots (1-31) for E1. This command associates the specified timeslots as ISDN channels. If you do not specify timeslots, all B-channels are associated to the PRI group. It is important to understand that the D-channel is always added to the pri-group, even if it is not explicitly specified, because it is required for signaling. Further, you can specify only one PRI group on the controller because each circuit has a single D-channel available.

After you specify the PRI group, the gateway automatically creates an associated voice port and serial interface. You configure the voice port in a similar way to the FXO voice port, as shown earlier.

The serial interface actually controls the D-channel (signaling channel). You do all configuration that affects processing of the signaling on the D-channel. It is on the D-channel that you can set the B-channel selection order if necessary. You do this by entering interface configuration mode on the D-channel serial interface and using the isdn bchan-number-order {ascending | descending} command.

After you have completed these steps, your service should be operational. Remember that during call setup, both the calling and called numbers are signaled on the D-channel. This allows call control to match on either field to route the call. See Chapter 9, "Dial Plans," and Chapter 10, "Digit Manipulation," for more details. Example 6-12 shows a typical E1 ISDN PRI configuration. In this example, the first four B-channels are used for voice traffic.

Timeslot 16 is the E1 D-channel that was added to the PRI group automatically. The B-channel selection order was also changed to begin with B-channel 1. Interface Serial1/0:15 and voice port 1/0:15 were automatically created when the PRI group was added to the E1 controller.

MGCP

If MGCP is going to control the ISDN PRI, a few configuration changes are required. With MGCP, the Layer 3 Q.931 messages are backhauled to the CallManager and not processed by the gateway.

When setting up the PRI group on the controller, you need to specify that MGCP will control the signaling. You do this from controller configuration mode by using the pri-group timeslots number service mgcp command.

Under the serial interface created for the D-channel, you also must code the isdn bind-l3 ccm-manager command. This attaches the D-channel Q.931 protocol to the CallManager backhaul. Example 6-13 shows the differences in configuration for an ISDN PRI under MGCP control.

The isdn bind-l3 ccm-manager command shown under interface Serial1/0:15 is available only after you add the service mgcp keywords to the PRI group definition. If the command is unknown when you try to configure it, look at the controller to ensure that the PRI group is defined correctly.

Verifying and Troubleshooting

To verify that the circuit is operating properly, you can use the show isdn status command. Example 6-14 shows the output of this command.

The example shows the status of the circuit at Layers 1, 2, and 3. Layer 1 should show ACTIVE, with the Layer 2 state as MULTIPLE FRAME ESTABLISHED. If that is not the case, check the physical interface for errors, as shown in the previous section. Also verify that the switch type is correct and that it matches the PSTN specification.

If Layer 2 is showing an incorrect state, you can use the debug isdn q921 trace to help determine the cause. Example 6-15 shows a problem with Layer 2, where both sides of the connection are configured as network, and the Layer 2 connection cannot be established.

If the status of Layer 1 and 2 are good and you are still having trouble placing calls, make a test call with debug isdn q931 active. This debug shows activity on the PRI D-channel and is useful in troubleshooting. First, if your test call produces trace output, it confirms that the request is actually being sent out the PRI interface. Second, as you can see in Example 6-16, it shows a great deal of information about the call. You can see the calling and called numbers, which B-channel is requested, and failure cause codes with reasonably clear descriptive text.

This example shows that extension 2001 called 1212 on B-channel 1. The call failed because 1212 is an unassigned numberthe number does not exist. The PSTN expects numbers in E.164 format. You can quickly see that the gateway is not passing numbers in the correct format to the PSTN, which is a dial plan configuration issue. You can also tell from this example that the B-channels are being selected in ascending order as was configured.

An additional command can provide assistance when you are troubleshooting CallManager MGCP-controlled trunks. It is the show ccm-manager command that, among other things, shows the status of the Q.931 backhaul. You can see this information at the end of Example 6-17. Other useful information provided by this command includes the destination IP address for the backhaul link, packet and error counts, and the slot and port information for the interfaces that are being backhauled.

Q.931 information is also written to the CallManager trace files and can be viewed with the Q.931 Translator utility provided with CallManager. This helps in troubleshooting by allowing you to look at historical call data. You can also use the Voice Log Translator (VLT) tool to parse and translate CallManager trace file data. It is available for download at Cisco.com.

Configuring E1 R2 Trunks

R2 signaling is a CAS method that is defined in International Telecommunications Union Telecommunication Standardization Sector (ITU-T) recommendations Q.400 through Q.490. This signaling method is widely used in Europe, Latin America, Australia, and Asia. You need to beware of many country-specific variants before you configure R2. These variants are described in the Consultative Committee for International Telegraph and Telephone (CCITT) R2 specification.

R2 signaling has two components: line signaling, which provides supervisory control signals, and interregister signaling, which provides call setup control.

Line signaling uses the ABCD bits in timeslot 16 for supervisory purposes, such as indicating line seizure or line clearing. Only the A and B bits are used in the CCITT-R2 format. Line signaling is supported in Cisco Voice Gateways in the following formats:

• R2-Digital Typically used for pulse code modulation (PCM) systems, this type is defined in ITU-U Q.421.
• R2-Analog Typically used in carrier systems, this type is defined in ITU-U Q.411.
• R2-Pulse This type is usually reserved for systems that employ satellite links.

Interregister signaling uses in-band multifrequency signals in each timeslot to send the calling number, called number, and call category. Three types of interregister signaling are supported:

• R2-Compelled In compelled, a tone-pair is sent from the switch and stays on until the remote end responds with a pair of tones that signals the switch to turn off the tones.
• R2-Noncompelled In noncompelled, the tone-pairs are sent as pulses so that they stay on for a short duration. Responses to the switch are sent as pulses.
• R2-Semicompelled In semicompelled, forward tone-pairs are sent as compelled. Responses are sent as pulses.

To configure an R2 signaling on an E1 trunk, you need to gather the following information from the PSTN service provider:

• Line signaling type.
• Interregister signaling type. Most implementations use R2-Noncompelled.
• Whether or not ANI information will be sent on the trunk.
• Number of channels, and which specific channels will be used to carry voice on the E1.
• Any CCITT-R2 country-specific variations that the PSTN uses.

To begin, you need to configure the physical port, as shown previously in the section "Configuring E1/T1 Physical Layer Connections."

The next step is to set up the DS0 groups. A DS0 group is a logical grouping of individual 64-Kbps voice channels on the E1. You can configure the entire E1 as a single DS0 group, or you can set up several DS0 groups, depending on requirements. For a PSTN connection, the DS0 logical configuration must match what the PSTN provides.

You configure the DS0 groups by using the following command:

ds0-group group number timeslots 1-31 type line-signaling-type interregister- signaling-type ani

The group number assigns a logical grouping to the channels defined in this DS0 group. You can specify up to 31 separate groups on an E1 trunk. Each logical DS0 group will have a separate voice port.

The timeslots keyword identifies the channels on the E1 that are being grouped under this DS0 group.

The line-signaling-type and interregister-signaling-type must match those that the PSTN service provider requires.

The ani optional keyword is added if the PSTN is providing ANI (calling number) information. The DNIS (called number) information is always provided.

Example 6-18 shows a typical E1 R2 configuration for a voice trunk in the United Kingdom. In this example, the first four channels of the E1 are used for voice. ANI information is being provided. Voice port 1/0:0 was created for the logical DS0 group 0. The cptone GB command under the voice port sets the locale to the United Kingdom.

Many country-specific variations of R2 signaling exist. To customize the voice gateway for the country where it is deployed, use the cas-custom channel number subcommand in controller configuration mode. The channel number parameter must match the group number of the DS0 group that you are customizing. Whenever you are setting up customized R2, Cisco recommends that you begin by using the country country use-defaults subcommand of the cas-custom command. This subcommand automatically sets up typical parameters for the chosen country. You can customize further from that starting point. Example 6-19 shows how you can use the cas-custom command.

In Example 6-19, the router generated the three subcommands following cas-custom 0. The actual command entered under cas-custom 0 was country hongkong-china use-defaults.

Note

You can find further information on country-specific R2 protocol customization on Cisco.com:

E1 R2 customization with the cas-custom commandwww.cisco.com/en/US/tech/tk652/tk653/technologies_tech_note09186a00800942f2.shtml

E1 R2 signaling configuration and troubleshootingwww.cisco.com/en/US/partner/tech/tk652/tk653/technologies_configuration_example09186a00800ad389.shtml

MGCP

CallManager MGCP does not support E1 CAS (E1 R2) trunks.

Verifying and Troubleshooting

You can view the status and signaling of the voice ports by individual timeslot (channel) using the show voice port summary command, as shown in Example 6-20. This example reflects the configuration shown previously in Example 6-18. Voice port 1/0:0 is controlling associated timeslots 1 through 4 with r2-digital signaling. You can also see that channel 1 is active (seized).

You can find information about the progress of a call on the voice port with the show voice call summary command. In Example 6-21, you can see that a call is in progress on voice port 1/0:0 channel 4.

If you are having problems completing calls on the E1 trunk, the debug vpm signal and debug vtsp all commands can provide useful information to assist in troubleshooting.

Configuring T1 CAS Trunks

T1 CAS circuits use in-band signaling, "robbing" the least significant bits from channels that carry voice to handle framing and to pass state information. The framing and state information is carried along with the voice information in each channel. This is in contrast to ISDN, which has one channel dedicated to carrying signaling for all the other voice channels.

To configure a T1 CAS trunk, you need to gather the following information from the PSTN service provider:

• Signaling interface FXO, FXS, or E&M. You also need to know the specific method that the PSTN uses to signal state of the line:

  - For FXO or FXSLoop-start or ground-start.

  - For E&MImmediate-start, delay-dial, wink-start (also known as Feature Group B (FGB)), double-wink-start (also known as Feature Group D [FGD]), or Feature Group D Exchange Access North American (FGD-EANA).

• Number of channels used for voice traffic If you are not using the full T1 for voice, you need to know which specific channels carry voice traffic.

Signaling information is passed over T1 circuits by emulating the methods used for analog trunks: FXO, FXS, and E&M. This information is passed using the A and B bits in SF circuits or the ABCD bits in ESF circuits. For more details on these signaling methods, see Chapter 5.

Loop-start signaling is the simplest CAS signaling method. Unfortunately, it has several disadvantages that make it an undesirable choice. It has no far-end answer or disconnect supervision. That is, you cannot relay when the remote side of the call answers or hangs up. Loop-start also provides no seizure of the channel on incoming calls. This can lead to a condition known as glare when both sides of the connection try to place a call at the same time.

Ground-start signaling has some advantages over loop-start. Ground-start can recognize far-end disconnect. It also can seize the channel on an inbound call, preventing the occurrence of glare. For this reason, ground-start is often used between PBXs or from a PBX to a voice gateway.

E&M signaling has many advantages over the other types shown so far and is the preferred choice for CAS trunks. E&M provides both answer and disconnect supervision and glare avoidance, can receive ANI information (FGD only), and can send ANI information (FGD-EANA only).

To begin, it is necessary to configure the physical port, as shown previously in the section, "Configuring E1/T1 Physical Layer Connections."

The next step is to configure DS0 groups. You configure DS0 groups in controller configuration mode using the ds0-group keyword. The syntax is ds0-group group-number timeslots 023 type signaling-type. The group-number parameter identifies the DS0 group and is a number in the range of 023. You can configure up to 24 DS0 groups, and you must assign each a different number. The timeslots 023 parameter establishes which channels in the T1 belong to this logical voice port. Table 6-3 lists the signaling types supported in current Cisco IOS versions that you might need for a PSTN connection.

It is important to note that if CallManager MGCP call control is used with the T1 CAS port, only E&M wink-start and E&M delay-dial are supported.

After you define the DS0 group, the gateway creates a voice port in the form of voice-port slot/port:ds0-group-number, as shown in Example 6-22. In this example, two DS0 groups were created. The first is associated with voice port 1/0:0, which controls timeslots 1 through 4 on the T1 and uses E&M wink-start signaling. The second is associated with voice port 1/0:1, which controls timeslots 5 through 8 on the T1 using E&M FGD signaling. Call control references the voice ports individually to route calls in and out of the associated timeslots on the T1.

MGCP

When you configure a T1 CAS connection that CallManager MGCP will control, you set all the call control parameters on the CallManager, rather than on the gateway. On the gateway, you associate the voice port created to the MGCP application with a dial peer.

CallManager MGCP supports only E&M wink-start and E&M delay dial signaling types for T1 CAS circuits.

You can control whether a port is used for inbound calls, outbound calls, or both.

You can also set the number of significant digits that CallManager should collect for an inbound call. You might need to do this if you are using DID and the last digits of the dialed number represent the extension where the call is to be routed. By setting the number of significant digits to the extension length, you can remove the need to further manipulate the incoming dialed number to deliver the call.

You can individually apply parameters to voice timeslots that are defined on the T1 controller.

Verifying and Troubleshooting

You can view the status and signaling of the voice ports by individual timeslot (channel) using the show voice port summary command, as demonstrated in Example 6-23. This example reflects the configuration shown previously in Example 6-22. Voice port 1/0:0 is controlling associated timeslots 1 through 4 with E&M wink-start signaling. You can also see that channel 1 is active (seized). Voice port 1/0:1 is controlling associated timeslots 5 through 8 with E&M FGD signaling.

You can also find information about the progress of a call on the voice port with the show voice call summary command. In Example 6-24, you can see that a call is in progress on voice port 1/0:0 channel 3.

If you are having a problem with calls on the port, you can use the debug voip vtsp all command, which shows the progress of the call control signaling for the call in progress on the port.

Configuring ISDN BRI Trunks

ISDN BRI the only service described that is not delivered on a T1 or E1 circuit. The information and tasks described in the section "Configuring E1/T1 Physical Layer Connections" do not apply to a BRI. BRI voice trunks are not used often in the United States; however, they are popular in other areas of the world.

An ISDN BRI is similar to a PRI in that the signaling information is carried separately from the voice-bearer traffic in a dedicated data channel. A BRI circuit has just two voice-bearer channels and one data channel (2B+D). Only two voice calls can occur simultaneously on a BRI.

BRI trunks operate in a master-slave mode. The master side is called the network termination (NT) side and is responsible for providing clocking and power. The slave side is called the terminal equipment (TE) side. This master-slave relationship extends to Layer 2 and Layer 3 protocols, as defined by ITU-T specifications Q.921 and Q.931, respectively. The master side initiates the Layer 2 and Layer 3 communications. When connecting to the PSTN, the voice gateway is always configured as the slave side (TE). An NT1 device is typically required to connect the voice gateway TE interface to the PSTN.

To configure an ISDN BRI voice trunk, you need to gather the following information:

• ISDN switch type Just as in PRI service, you must specify the switch type in use for signaling to operate properly between the voice gateway and the PSTN. You must globally configure an ISDN switch type on the gateway. You can override this global definition on a port-by-port basis if necessary.
• ISDN service profile identifiers (SPID) if required The PSTN assigns ISDN SPIDs and local directory numbers (LDNs). These are only used for national and DMS-100 switch types.

Begin configuration by defining the ISDN switch type in global configuration mode using the isdn switch-type switch-type command. Table 6-4 lists ISDN switch types supported by current versions of Cisco IOS that you might encounter when setting up a PSTN BRI trunk. You can override this setting at the port level if necessary to support multiple connections to different switch types.

Continue the configuration in interface configuration mode using the interface bri slot/port command. If necessary, configure the first SPID using the isdn spid1 spid [ldn] command, as well as using the PSTN-assigned SPID and LDN numbers. You can configure the second SPID in the same manner using the isdn spid2 spid [ldn] command.

For voice calls to be properly routed to digital signal processor (DSP) resources for voice processing, you must also code the isdn incoming-voice voice command on the interface.

Example 6-25 shows the configuration of a typical ISDN BRI voice trunk.

MGCP

CallManager versions 4.1.3SR1 and later allow you to control BRI voice ports with MGCP. With MGCP, the Layer 3 Q.931 messages are backhauled to the CallManager and are not processed by the gateway.

Under the BRI interface, you must code isdn bind-l3 ccm-manager service mgcp. This attaches the D-channel Q.931 protocol to the CallManager backhaul. Example 6-26 shows the configuration for an ISDN BRI under MGCP control.

In Example 6-26, the ISDN switch type is basic-net3. This is the only switch type supported by CallManager for MGCP control of a BRI.

Verifying and Troubleshooting

To verify that the circuit is operating properly, you can use the show isdn status command. Example 6-27 shows the output of this command. The example shows that Layers 1 and 2 are active and that no Layer 3 calls are currently in progress. You can also see the ISDN switch type configured globally and for the specific port.

If you have not properly established the Layer 2 connection, use the debug isdn q921 trace to help identify the problem. If the Layer 1 and Layer 2 output looks correct and you are still having trouble placing or receiving calls on the port, use the debug isdn q931 trace command. Both of these are described in the troubleshooting portion of the "Configuring ISDN PRI Trunks" section.