6.1 Introduction

6.2 Default Channel Assignment

6.3 Exclusive Mode

6.4 Non-Exclusive Mode (Preferred Mode)

6.4.1 Switching Considerations For Disconnect Handling

6.5 Assigning Incoming Calls to TCP Instances

6.1 Introduction

When a TCP instance is launched, a B channel is assigned to it. This channel is called the default channel. When placing a call, the TCP will always request the use of the default channel from the network. It directs the network to respond in one of two ways, or modes:

· Exclusive Mode: If the TCP's default B channel is not available on the network, the network clears the call.

· Non-Exclusive Mode (also called Preferred Mode): If a TCP's default B channel is not available on the network, the network directs the TCP to use another channel. In this case, the application must perform switching, to connect the new B channel with the DSP resource associated with the TCP.

This chapter describes how a TCP instance is assigned a default channel. It also describes Exclusive Mode and Non-Exclusive Mode, and explains how to set the mode for each TCP instance.

Note: In the US, ISDN systems almost always place and receive calls in Exclusive Mode. In Europe, both modes are used: many central office (CO) switches require the TE equipment side to use Non-Exclusive Mode.

For details about the switch model for a board, see the installation and developer's manual for the board.

6.2 Default Channel Assignment

A default channel is assigned to each TCP instance in either of the following ways:

· Ordinarily, the network stream and timeslot are inferred from the DSP stream and timeslot that were specified when the context was opened. This is the trunk slot that is connected to the DSP slot when the EnableMvip line in the configuration file is set to NO. (See Section 3.3, Making Switch Connections for NMS ISDN for more information.)

· The default channel can be specified by setting the following parameters for that TCP instance (these parameters are documented in Appendix A):

Service	Parameters
NCC	NCC.ADI_ISDN.START_EXT.networkstream NCC.ADI_ISDN.START_EXT.networkslot
ADI	ADI.ISD.networkstream ADI.ISD.networkslot

For example, MVIP-95 0:0, 1:0¹ on an AG Quad T (Trunk A) corresponds to B channel 1.

6.3 Exclusive Mode

In Exclusive Mode, a TCP instance effectively owns its default channel. To place an outbound call, the TCP asks the network for the default channel. If the channel is not available, the network clears the call. When the TCP senses this, Natural Access sends xxxEVN_CALL_DISCONNECTED to the application.

For example, suppose the application attempts to establish an outbound call using a TCP which is set up to use default B channel 5 in Exclusive Mode (corresponding to MVIP-95 0:4, 1:4¹). The TCP asks the network for B channel 5. If the channel is available, the network signals the TCP to continue with the call on that channel. Otherwise, the network clears the call.

In Exclusive Mode, any incoming call will be routed to a TCP instance only if the call is on its default channel. All calls on that channel will be routed to the TCP.

To set a TCP instance to Exclusive Mode, set the NCC.ADI_ISDN.START_EXT.exclusive parameter to 1. (Under ADI, this parameter is ADI.ISD.exclusive.)

You can also use the NCC.ADI_ISDN.START_EXT.direction parameter to specify whether the channel is incoming-only, outgoing-only, or bidirectional. Incoming calls will never be offered to an outgoing-only TCP channel. A call placed on an incoming-only channel will be rejected. (Under ADI, this parameter is ADI.ISD.direction.)

Note: The setting of the direction parameter should match the configuration of the connected equipment (for example, the network switch). For example, if the application attempts to place a call on a channel that is configured on the connected equipment to be incoming-only, then the outbound call will always be rejected.

6.4 Non-Exclusive Mode (Preferred Mode)

When a TCP instance places a call in Non-Exclusive Mode, it requests the use of its default channel from the network. The network may allow the use of the default channel, or direct the TCP to use a different channel. If a different channel is assigned, the application must perform switching to route the call to the DSP resource associated with the TCP instance.

To do so, the application must:

1. Determine which B channel is chosen by the network for the TCP (for example, the default channel or another channel). If no channel is available, the call is rejected by the network.

When the application receives the xxxEVN_PLACING_CALL event, the stream and timeslot corresponding to the channel are available in a call status structure:

Service	Structure	Fields	To access, invoke...
NCC	NCC_ISDN_EXT_CALL_STATUS	stream timeslot	nccGetExtendedCallStatus
ADI	ADI_CALL_STATUS	stream timeslot	adiGetCallStatus

2. Connect the stream and timeslot for that channel with the stream and timeslot assigned to the context on which the call is placed.

To perform this connection, the application can use the Natural Access Switching service.

The switching should be performed as soon as possible after xxxEVN_PLACING_CALL is received, to ensure that call progress analysis succeeds.

For examples of applications that perform this switching, see the isdnchan and isdnncc sample programs documented in Chapter 7.

For example, a TCP attempts to establish an outbound call on its default B channel. In this case, the default channel is 5, corresponding to MVIP-95 0:4,1:4. The network indicates that channel 5 is not available, but channel 9 is available (corresponding to MVIP-95 0:8,1:8). Since the TCP is running in Non-Exclusive Mode, the call is established on channel 9. The application must now connect the TCP instance's DSP resource (MVIP-95 16:4, 17:4) with the B channel to establish the voice path (for example: 0:8 => 17:4, 16:4 => 1:8)².

Incoming calls from any B channel may be routed to the TCP instance unless the B channel on which the call is located is assigned to another TCP running in Exclusive Mode. For details, see Section 6.4.1, Switching Considerations For Disconnect Handling.

To set a TCP instance to Exclusive Mode, set the NCC.ADI_ISDN.START_EXT.exclusive parameter to 1. (Under ADI, this parameter is ADI.ISD.exclusive.)

6.4.1 Switching Considerations For Disconnect Handling

(Non-Exclusive Mode only) When a call is disconnected, the application must disconnect only the trunk-to-DSP side of the full duplex switching connection that it made when the call was established. The DSP-to-trunk connection must not be disconnected because it may interfere with a new call arriving on the B channel that is handled by another thread or process.

Disconnecting the trunk-to-DSP connection ensures that if a new outgoing call is placed on the context, any call progress analysis functions will receive only silence until the application makes a new connection.

6.5 Assigning Incoming Calls to TCP Instances

When a call arrives on a B channel, a TCP instance is assigned to the channel as follows:

1. If a TCP instance is currently assigned to the channel in Exclusive Mode (and is not set up as outgoing-only), the call is routed to the TCP.

If the TCP instance is outgoing-only, the call is rejected.

2. If a TCP is currently assigned to the channel in Non-Exclusive Mode (for example, the channel is the TCP's default channel), the call is routed to the TCP if it is not already handling another call (and is not set up as outgoing-only).

3. Otherwise, the call is passed to any free TCP that is in Non-Exclusive Mode and is not set up as outgoing-only.

4. If no TCP is available, the call is rejected.

In MVIP-90 terms, this connection is 16:8 <=> 18:4.