(Page 1 of 1 in this chapter) Version


Appendix C

NFAS and D Channel Backup


Introduction
Non-Facility Associated Signaling (NFAS)
Setting Up NFAS
Using NFAS
D Channel Backup
Setting up D Channel Backup
Handling D Channel Failure

Introduction

In setups with multiple T1 ISDN trunks, you can set up a single D channel to serve all trunks. This configuration is called Non-Facility Associated Signaling (NFAS). This appendix describes NFAS. It also describes how to create a secondary D channel, for use when the primary channel goes out of service.

Non-Facility Associated Signaling (NFAS)

NFAS is supported on T1 trunks.

Data on a T1 trunk is transmitted in channels. For primary rate ISDN, T1 carries 24 channels. With primary-rate ISDN:

In an NFAS configuration, trunks are grouped into one or more NFAS groups. One of the trunks in each group has a D channel carrying the signaling for all of the B channels on all of the trunks in the group. This leaves channel 24 free on all other trunks in the NFAS group. This extra channel can be used as another B channel. (See Figure 32.)

Figure 32. Sample NFAS Configuration


A single NFAS group can contain trunks from multiple boards. (See Figure 33.)

Figure 33. Sample NFAS Configuration (Multiple Boards)


Note:  If your application will use the AG ISDN software in Channelized configuration, all trunks in an NFAS group must be on the same board; an NFAS group cannot contain trunks from multiple boards. This restriction does not apply to the ACU configuration or the LAPD configuration.

Setting Up NFAS

NFAS groups are specified in the AG configuration file. In the file, you specify which trunks on which boards belong to which groups. You also specify which trunk in the group will carry the D channel. All other trunks are set to DigitalMode=RAW, so all 24 channels on these trunks can be used as D channels. For more information, see the AG ISDN Installation Manual.

Your application initializes just as described in Chapter 3. The only difference is that when the application calls isdnStartProtocol, the NAI specified in the call can be greater than 3.

Using NFAS

The fact that a given trunk is a part of an NFAS group, and the D channel resides elsewhere, is largely transparent to the application. There are two slight behavioral differences:

D Channel Backup

When NFAS is used, the reliability of the signaling performance for the ISDN interfaces controlled by the D channel may be improved by creating a standby
D channel: the D channel backup. The D channel backup feature allows continued access to the ISDN network if one of the D channels fails, by transferring most of the signaling information to the backup D channel.

The designated primary D channel is always present on one trunk. On a different trunk, a backup or standby D channel is present (see Figure 1).

Figure 34. D Channel Backup


At any point in time, only one of the D channels conveys B channel signaling information. The other D channel remains in a standby role and is active at the LAPD layer (layer 2) only. While the backup D channel is on standby, any layer 3 messages received on it are ignored.

Neither D channel can serve as a B channel while designated as a backup D channel. Also, each D channel pair provides signaling only for the set of B channels assigned to it, and cannot backup any other D channel(s) on a different interface.

When both D channels are out of service, the first D channel has priority as the channel to carry call control signaling. If the first D channel cannot be established, then the backup D channel is chosen.

Setting up D Channel Backup

The D channel to use as the backup is specified in the AG configuration file. For more information, see the AG ISDN Installation Manual.

Note: Both primary and backup D channels must be defined on the same board and belong to the same NFAS group. Set DIGITALMODE to PRI for both.

When starting ISDN with CT Access, you must enable option t309 in the ISDN_PROTOCOL_PARMS_Q931CC or ISDN_PROTOCOL_PARMS_CHANNELIZED structure, which appears as:

WORD t309; /* T309 in use flag */

To enable this option, set this parameter to 1 in your application. When enabled, active calls are preserved while switching to the backup D channel when the primary D channel fails. For more information, refer to the AG ISDN Messaging API Developer's Reference Manual.

Handling D Channel Failure

When the primary D channel fails, B channel signaling information carried by the channel is transferred to the backup D channel. When a transition occurs, most stable calls, i.e., those calls which are in the Active, or Connected, states, may be preserved, although Message-Associated User-to-User Information (MA-UUI) and both call-associated and non-call-associated Temporary Signaling Connections, may be lost. There is a small interval (controlled by timer t309) after a failure of the LAPD link before the B channels are removed from service.

If the primary D channel fails and timer t309 is enabled (set to 1), any calls which are were in the Connected state at the time of the failure will be preserved. Any calls which were initiated but had not entered the Connected state are cleared. Each cleared call receives xxxEVN_CALL_DISCONNECTED.

If the primary D channel fails and timer t309 is disabled (set to 0), all initiated calls on both terminal and network sides are cleared with xxxEVN_CALL_DISCONNECTED messages.

After the data link is reestablished on the backup D channel, the application can start placing and receiving new calls. If the data link on the backup D channel cannot be established, the stack will keep trying to establish the link until one of the D channels comes up.

A situation may arise when the primary D channel fails, and one side of a call in Connected state initiates the disconnect process (leaves the Connected state) just before the data link failure is detected in the system. At the moment of failure recognition, one side will be still in the Connected state and the side which started the disconnect sequence will not be in the Connected state anymore. In this case, the state of the first side is preserved. The second side receives an xxxEVN_CALL_DISCONNECTED message. When the data link is reestablished, the first side will still be in the Connected state. The application should disconnect and release this call.

A similar situation may arise when at the moment of the primary D channel failure both sides are in the Connected state and their states are preserved, but then one side initiates the disconnect process and is cleared before the data link is reestablished. As a result, when the data link is reestablished, one side will still be in the Connected state and will not know that the other side has disconnected. As in the case described above, the application should sense this situation and disconnect and release the connected call.

Since a channel cannot serve as a B channel while designated as a backup D channel, an application is not allowed to place a call on a timeslot on a primary or backup D channel. Both primary and backup channels are assigned to the slot number 24 on their corresponding trunks. Thus, the maximum B channel slot number for a trunk with a primary or a backup D channel is 23. If the application attempts to place a call on a D channel timeslot, the application receives xxxEVN_CALL_DISCONNECTED with the following reason codes:
Function

Reason Code

ADI_CALL_DISCONNECTED

ADI_DIS_SIGNAL

NCC_CALL_DISCONNECTED

NCC_DIS_SIGNAL



(Page 1 of 1 in this chapter) Version


tech_support@nmss.com
Copyright © 1999, Natural MicroSystems, Inc. All rights reserved.