[ Previous | Next | Contents | Search ]
AIXLink/X.25 1.1 for AIX: Guide and Reference

Configuring X.25 Communications with SMIT

The easiest way to configure the X.25 software is through SMIT. When you are configuring the system, options are presented based on what adapters are available in the system.

Note: It is best if all the adapters to be used are installed in the system before the X.25 configuration is begun. Adapters can be added or deleted at a later time, but for the instance numbers to be kept in a simple order, as much of the configuration as possible should be done at one time.

The main areas of configuration are:

From a system where the adapters are installed, the following steps outline those necessary to configure X.25 ports ready for use:

  1. If AIX Version 4.1 or later is installed, configure the total number of virtual circuits allowed on the system. This is specified on the X.25 license.
  2. Add device drivers to each adapter to be used. If you use the IBM 2-Port Multiprotocol PCI Adapter, you must add the hdlc device driver. Otherwise, you must add the twd device driver.
  3. Add X.25 ports to each of the adapter's physical ports.
  4. Change the configuration parameters on any ports where the settings must be different from the defaults.
  5. Add COMIO emulation to any of the X.25 ports where this is required. Refer to "Managing COMIO Emulation".
  6. Add TCP/IP support to any of the X.25 ports where this is required. Refer to "Managing TCP/IP Configuration".
  7. Start Triple-X PAD if it is used on the system. Refer to "Managing the Triple-X PAD".

Initial SMIT Path

The X.25 software is configured based on the adapter it is to run on. To reach the communication devices screen:

  1. Enter: smit
  2. Select Devices.
  3. Select Communication.

    Under the communication devices list are entries for the adapters supported in the system. Select the adapter type required and a list of the available software packages for that adapter type is listed. Depending on the adapter, select one of the following:

  4. Select Adapter.
  5. Select Manage Device Drivers for the adapter type selected.
  6. Select Manage X.25 "Licensed Progam" Product Device Driver."
Note: After you select IBM 2-Port Multiprotocol PCI Adapter do the following:
  1. Select Manage HDLC Device Driver
  2. Select Manage Additional Protocols / Emulators
  3. Select Manage X.25 over HDLC Device Driver

Managing Device Drivers

The screen shown in the following figure allows the instances of the twd device driver to be managed. If a driver is in the defined state, Configure a Defined Device Driver makes it available, assuming that no other driver is configured at the adapter. The device driver is added to an adapter before the X.25 ports on that adapter are added, and is removed after all the X.25 ports on the adapter are removed.

Managing Ports

Select Manage X.25 Ports to manage the X.25 ports on the system, from the Licensed Program's initial SMIT screen. This brings up the Manage X.25 Ports SMIT screen as shown in the following figure.

From the Manage X.25 Ports SMIT screen, the following actions can be taken:

List All Defined Ports List all the ports configured on the system, either in the defined or available state.
Add a Port Add a new port definition. This option sets all defaults for the port depending on the country prefix you choose. If you do not choose a country prefix, port defaults are set to a universal setting.
Move a Port Definition The definition of a given port can be moved to a different port, possibly on a different adapter.
  1. Select the port to be moved.
Note: For the IBM 2-Port Multiprotocol adapter, only those ports currently in the defined state may be moved.
  • Select the new parent adapter driver.
  • Enter the new port number.
  • Note: For the IBM 2-Port Multiprotocol adapter, the port number is automatically filled in for you.
    Change/Show Characteristics of a Port The characteristics of the port are controlled by a number of attributes, divided up in SMIT into a number of screens. This section gives access to those attribute screens.
    Remove a Port The port is removed from the available state back to the defined state. When a port is in the defined state, the port's configuration information is kept in the ODM database. Therefore, the port can be reconfigured into the available state. An option is given to also remove the port's information from the ODM database. If this option is chosen, the port is completely removed from the system.
    Configure a Defined Port If a port is in the defined state, then this selection makes it available.
    Add COMIO Interface to a Port Use of the emulation of the base AIX Version 3 support is optional. It would be needed if applications written to the base AIX Version 3 APIs such as xtalk or SNA are used. If this is required on a given port, then COMIO should be added to produce a new device entry such as /dev/x25s0 . The emulator can be added to any available port. No additional configuration information is required.
    Remove COMIO Interface to a Port The emulation of the base AIX Version 3 support for a given port can be removed.
    Add TCP/IP Interface to a Port TCP/IP can be enabled on any of the available ports. Details of the port's TCP/IP information is then provided in a separate screen.
    Remove TCP/IP Interface to a Port The TCP/IP support for a given port can be removed.
    Manage PAD If the Triple-X PAD support is to be used on the system, then the PAD software should be configured.

    Adding a Port

    To add a port, select Add Port and a list of the available adapters will be presented. When you select an adapter, a SMIT screen will allow entry of the port's NUA. Information on the attribute choices available is given in the help text. It is probable that the port will require further configuration before use to match its characteristics with those provided by the network provider. Use the Change/Show Characteristics of Port screen to alter the attributes.

    The following attributes must be entered:

    PORT number The physical port on the adapter's expansion cable that is to be used. With the Portmaster adapters, this can vary between 0 and 7 for the V.24/RS-232 interface and between 0 and 5 for its other interfaces. For the X.25 Co-Processor adapters, this must be 0. If the port is associated with an IBM 2-Port Multiprotocol adapter, the port number is automatically filled in, and does not require an entry.
    Local network user address The Network User Address for the port. This is usually supplied by the network provider and must agree with the subscription obtained.

    The following attributes are optional. See "General Parameters" for more details.

    Network Identifier On networks that implement X.121 addressing, the country prefix uniquely identifies the country. In some cases, this can be used to identify default characteristics of the network. In some countries, there is more than one network type, and though they share the same country code, they have differing characteristics. If the identification for the local network is listed, then that should be selected. Otherwise, select other public if attached to a standard network. Select other private if the network's addressing scheme is different from X.121.
    Country prefix The country code provided by X.121. This should be made available by the network provider.
    VC ranges For each of the four types of virtual circuit, the initial channel number, and the number of channels must be given. This must match the configuration of the local DCE, and so should be based on the information provided by the network provider.

    Moving a Port

    The attributes defining a port can be transferred to a different port. When Move Port Definition is selected:

    1. Choose existing port to be moved.
    2. Select the new adapter to be used. The new adapter can be the same as the port's current adapter.
    3. Enter the physical port number on the given adapter.

    Configuring a Port

    From the "Manage X.25 Ports" SMIT screen, select Change/Show Characteristics of Port. The attributes for the port are divided into a number of categories. They are listed in the SMIT screen in the order in which they will most likely be used (see the Change/Show Characteristics of Port SMIT Panel figure). It is often not necessary to configure anything other than the port's general parameters.

    General Parameters

    General parameters are those most likely to need configuring. For a given port, there are the following attributes:

    NUA The network user address for the port. This is usually supplied by the network provider and must agree with the subscription obtained.
    Calling address On a few networks, the NUA of the machine originating the call cannot be put into the call it generates. Usually the calling address is allowed. This attribute is used only with COMIO connections.
    Enable DLPI If direct access to the frame layer is required on this port, then DLPI should be enabled. This disables packet layer (X.25) access. DLPI is usually disabled.
    VC ranges VC ranges are constructed according to the lowest logical channel number and the logical channels desired for each of the VC types. The lowest logical channel number indicates the lowest-numbered logical channel that may be used for a specific VC type. This number depends on the network subscription and must be in the range of 1 through 4096. The lowest logical channel may be 0 only when the VC type is not used. Logical channel number 0 may not be used for a VC channel because it is always used as the network protocol channel. Check with your network provider to learn if logical channel 0 is included in a VC range.

    The number of logical channels indicates how many VCs to configure starting at the lowest logical number of that VC type. This information is supplied by the network provider.

    The specified VC channel ranges must not overlap and must not exceed the total number of channels allowed. In addition, the logical channel numbers must be defined according to the following rule:

    PVC < Incoming SVC < Two-way SVC < Out-going SVC

    If the range of VCs is not needed, set the lowest logical channel number and the number of logical channels for the VC type to 0.

    X.32 This set of attributes should be set if the network provider requires X.32 DTE identification. Both the X.32 XID identity and the X.32 XID signature are supplied by the network provider. For more information, refer to "Configuring a Port for X.32".
    Dial-Up These attributes must be set if a dial-up connection is used. For more information, refer to "Dial-Up Parameter Descriptions"
    Packet Parameters

    Packet parameters give the characteristics of the packet layer.

    CCITT The level of CCITT recommendation that is implemented by the network. The most significant changes are between 1980 and 1984 where packets sizes greater than 1024 were introduced, along with modulo 128. It should be selected to reflect the implementation of the network.
    Modulo The packet layer modulo must agree with that of the network. Usually, it is 8. This governs the number of packets that can be sent out before the sender must wait for acknowledgement of the data. On some networks, modulo 128 can be used. Though this would not normally be of benefit, the network provider would be able to advise on its usefulness in a given situation.
    Type of line By default, the system being configured will run as a DTE. In exceptional cases where two systems are being run back to back, one can be set up as a DCE. This does not allow the system to run as a full DCE; rather it just allows back-to-back operation.
    DTE/DCE alteration In cases where two systems are being run back to back and the remote system is also running with this licensed program, the selection of DTE/DCE can be made automatically. In cases where one system is a DTE connected to a network, this parameter need not be used.
    Registration Some networks support registration, where the DCE configures some of the DTE's characteristics automatically. This is not normally the case, and so this should usually be set to no.
    A bit On most X.25 networks, the method by which the addresses are coded into the call packet is standard. However, on some networks this differs. By far the more common is one called non-TDA/NZI the section addressing, and this is standard for most networks. The other mode of addressing is TOA/NPI. For regular usage, set the A-bit to off, which provides the regular non-TDA/NZI addressing. Set this to on only if it is required by the network provider.
    NIDU The data passed up from the packet layer to applications such as TCP/IP is broken up into data units. Typically, the size of the data unit should be such as to wholly contain the typical application data packet. For example, if the MTU for the X.25 TCP/IP interface is IS00, set the NIDU parameter to I500.
    Packet sizes These attributes only apply to switched virtual circuits used on the port. The default packet sizes are used when no negotiation of packet sizes takes place between the DTE and the DCE. Therefore the default sizes depend on the network subscribed to and should match the networks requirements. The default packet size does not have to match the remote systems default packet size unless the systems are running back to back. The maximum packet size is used if negotiation does take place between the DTE and DCE. The maximum size should be set to a size sufficient to contain the average size of the data unit used.
    Window sizes The default and maximum packet window size that can be used. These attributes specify how many packets can be sent before acknowledgement is received. The window sizes should be set according to the recommendations of the network provider. Generally, for slow, reliable communication links, a larger window size is recommended. For unreliable communication links, a small window size is desirable. When two systems are running back to back, their window sizes must match.
    Throughput class Throughput is the DTE-to-DTE transfer rate in bits per second (bps). The throughput can be affected by the local system, remote system, and the network. The default throughput class should be set to the line rate.
    T (timers) The packet layer timers, T20 through T28, implement timeouts for different types of packets.
    R (counters) If the particular timer packet is not confirmed within the given period, it is re-sent, based on the value of the equivalent R counter. If still no acknowledgement is received, a recovery process is started. The nature of the process started depends on which type of packet was left unacknowledged.
    Fast Select An optional facility that is available from some network providers. In a fast-select call, the call request is always rejected, but the CUD in the call and clear constitute the information exchange. This saves having to have the call established if only a small amount of data needs to be sent.
    D bit If the applications being run require DTE-to-DTE data acknowledgement, they will make use of the D bit. On some networks, this is a separate option, and so may not be subscribed to. If this is the case, then D bits should not be allowed, and this parameter should be set to forbid. Usually, they should be allowed.
    Reverse Charging An optional facility that allows a call to specify that the receiving party pays for the call. In all cases, allow allows it, while allow if not billed allows it only in outgoing calls.
    Facility size This gives the maximum number of bytes of facility that can be in a call.
    Interrupt size This gives the maximum number of bytes of data that can be held in an interrupt packet.
    Frame Parameters

    The frame parameters give the characteristics of the frame layer.

    T (timers) The T1 timer gives the timeout by which an acknowledgement of the data sent should have been received. On slow lines, especially with large data packets, the value of T1 should not be too low. If too low, the frame will not be fully transmitted before the timer expires, and so a high number of T1 expirations would be seen. N2 specifies, if T1 does timeout, how many times the frame should be retransmitted. T2 is less likely to be varied, specifying to the software the required turn around time for a frame. T3 and T4 ensure that any frame layer failure is recognized.
    Frame modulo Specifies whether the counters that check frame-layer activity between the DTE and the DCE work in modulo 8 or 128. Usually modulo8 is used. This value will be based on the network.
    Frame window This gives the number of frames that can be sent to the DCE before the DTE must wait for acknowledgement. The value, typically, does not affect performance greatly and should be based on the network providers' recommendation.
    DTE/DCE Selection on whether the system operates as DTE or DCE can be made at the packet or frame layers. This selection can be automatic (based on the ISO standard ISO8208) or fixed. However, this does not allow the system to act as a full DCE.
    Connection mode Once the physical layer is connected, the frame layer is established. The different networks require different modes of startup at the frame layer. One of the most common is for the DTE to start sending out SABM frames to indicate its frame layer is ready. The network provider will provide additional information in this area.
    Permanent Virtual Circuit (PVC) Parameters

    The characteristics of a PVC cannot be changed at the time of usage. Unlike SVCs, they have no call request/confirm exchange in which to negotiate packet size. The X.25 Default PVC Parameters SMIT screen allows the defaults that are to be assumed for this port to be set. If a specific PVC does not match these defaults, then this should be given as a non-default PVC. The Manage Non-Default PVC screen is provided for this. As identification of a given PVC is based on its logical channel number, this must be specified when dealing with non-default PVCs.

    Packet sizes The packet sizes to be used.
    Window sizes The packet window sizes to be used.
    D bit If the applications being run require DTE-to-DTE data acknowledgement, they make use of the D bit. On some networks, this is a separate option, and so may not be subscribed to. If this is the case, then D bits should not be allowed. Usually, D bits should be allowed.

    Configuring a Port for V.25bis Dialing

    The CCITT V.25bis Recommendation provides a protocol to connect automatic calling and/or answering equipment to a telephone network. This recommendation contains two types of connections. The connection type is distinguished by the DTR signal used. If the DTR signal is 108.1, the connection type is called direct. A direct V.25bis connection uses only V.24 signals to establish connections. If the DTR signal is 108.2, the connection type is called addressed. An addressed connection uses V.25bis commands and V.24 signals to establish connections. To use the addressed mode, your modem or ACU must support the V.25bis command set.

    The V.25bis functionality provided in the Streams X.25 LPP configures the port for both incoming and outgoing calls. Therefore, your port can both originate and answer calls without having to be reconfigured.

    Use the following method to configure a port for V.25bis dialing.

    Using SMIT, enter:

    smit

    Select Manage X.25 Ports.

    Select Change/Show Characteristics of Port.

    Select Change/Show X.25 General Parameters.

    On the Change/Show X.25 General Parameters screen, there is a Dial-Up Configuration menu that displays certain options for setting up calls.

    The setup in the SMIT menus must match the setup of the actual modem or autocall unit (ACU) you are using. Therefore, it is important that you are familiar with the modem setup before you attempt to configure the port for V.25bis.

    Here is a sample setup for using V25bis addressed mode. A phone number must be specified if the port is to originate any calls. When originating a call, the X.25 adapter sends the dial string included in the Phone Number or Address to Call field to the modem or ACU using V25bis commands.

    Dial-Up Configuration
    **********************
    Connection type                           [V25bis]        +
    V25bis Call Establishment Method          [Addressed]     +
    Phone Number or Address to Call           [5551234]
    Maximum Connection Delay                  [30]            #
    Enable/Disable DSR Polling                [enable]        +
    DSR polling timeout                       [30]

    Here is a sample setup for using V25bis direct mode:

    Dial-Up Configuration
    **********************
    Connection type                           [V25bis]         +
    V25bis Call Establishment Method          [Direct]         +
    Phone Number or Address to Call           []
    Maximum Connection Delay                  [30]             #
    Enable/Disable DSR Polling                [enable]         +
    DSR polling timeout                       [30]

    Dial-Up Parameter Descriptions

    Connection Type There are three types of connections:
    Direct/Leased Default connection type.
    V25bis Represents all dial-up connection types with the exception of manual dial.
    Manual-dial Connections where the user does the dialing instead of the modem.
    V25bis Call Establishment Methods There are two V25bis call establishment methods:
    ADDRESSED Uses V25bis commands to communicate with the modem or ACU. For originated calls, a dial string must be specified. This string is sent to the modem when the DTE requests the call. When answering calls, the modem uses V25bis commands to notify the DTE of an incoming call.
    DIRECT Uses V.24 signal control to establish connections. When originating a call, the mode must already know the dial string to call. When answering calls, the modem or ACU uses the Calling Indicator (circuit 125) to indicate an incoming call to the DTE.

    Addressed mode is used on outgoing calls when the DTE sends the number to dial to the modem. Addressed mode is used on incoming calls when the modem uses V25bis commands to notify the DTE of an incoming call.

    Direct mode is used on outgoing calls when the modem already knows the number to dial. Direct mode is used on incoming calls when the modem uses the Calling Indicator (circuit 125) to indicate an incoming call to the DTE.

    Phone Number or Address to Call Specifies the phone number the modem should call to establish a connection. The phone number may contain the digits 0 through 9 inclusive and any of the following dial modifiers:
    : Wait tone
    < Pause
    = Separator 3
    > Separator 4
    P Dialing to be continued in pulse mode
    T Dialing to be continued in DTMF mode
    & Flash

    These dial modifiers are only valid if using V25bis addressed mode.

    Since the port may not be used to originate calls, the phone number is not a required field. However, if you do not specify a phone number and you try to send a call using addressed mode, the call will fail.

    Maximum Connection Delay

    Maximum time in seconds to allow for connection establishment. Must be set long enough for modems to perform handshaking or training operations. When originating calls, this value should be smaller than the T3 idle timer frame-layer parameter.

    Used for all dial-up connections. For leased connections, this value remains unchanged.

    Enable/Disable DSR Polling DSR polling must be enabled if using a V25bis or manual-dial dial-up connection. If using a leased connection, DSR polling must be disabled.
    DSR Polling Timeout After call establishment, time in seconds to wait after DSR has dropped before ending the connection. Used only if DSR polling is enabled.

    Configuring a Port for X.32

    Use the following method to configure a port to use X.32 XID authentication.

    Using SMIT, enter:

    smit

    Select Change/Show Characteristics of Port.

    Select Change/Show X.25 General Parameters.

    On the Change/Show General Parameters screen, the settings for the X.32 attributes must be obtained from your network provider.

    X.32 Parameter Descriptions

    Use X.32 XID Exchange Indicates whether or not to use X.32 XID authentication procedures when the link is established.
    X.32 XID Identity Represents the X.32 identity of your port. This attribute must be provided by the DCE you are calling. The identity should be a string of 64 hexadecimal characters.
    X.32 Signature Identifies your password. This attribute must be provided by the DCE you are calling. The signature should be a string of 64 hexadecimal characters.

    Managing COMIO Emulation

    From the Manage X.25 Ports SMIT screen, emulation of the AIX Version 3 base X.25 driver support can be added to an X.25 port. This screen also allows it to be removed. For details on this screen, see "Managing Ports".

    When Add COMIO Interface to Port is selected, a choice of available ports is presented. Only one emulator can be added to any one port, and if it has already been added to the port chosen, the operation will fail.

    When Remove COMIO Interface from Port is selected, a choice of the emulators configured on the system is presented.

    Note: On AIX Version 3, the base X.25 support produces COMIO device-specific files with the same names as the X.25/6000 Licensed Program. The instance numbers of the ports are unique, but the two types of ports cannot be distinguished by their names. See the lsx25 command for details on how to distinguish the two types.

    Managing TCP/IP Configuration

    TCP/IP support can be added or removed from a given port. For more details of the TCP/IP support, refer to the section on TCP/IP in the AIX Version 4.3 System Management Guide: Communications and Networks.

    Add a TCP/IP Interface

    To add TCP/IP support, select Add TCP/IP Interface to Port from the Manage X.25 Ports screen and select the correct adapter. The following Add an X.25 "Licensed Program" Product Network Interface screen displays:

    Internet Address The IP address to be given to this interface, for example, 192.35.231.224. There should be a host name for this address in /etc/hosts.
    Network Mask The netmask of the network depends on the network's implementation. If the IP connection is simply from this port to another, then the default, which is obtained by leaving the field blank, is usually acceptable. If many X.25/IP connections are being used to make up a network, the netmask should be consistent across the systems.
    "Licensed Program" Product Port The X.25 port selected.
    Activate Choose whether to immediately activate the interface or not. If there is still configuration work to be done on the port, then it should be activated later with the ifconfig command.

    Remove a TCP/IP Interface

    To remove a TCP/IP interface from a port, select Remove TCP/IP Interface from Port and then choose the interface to be deleted.

    Note: Use ifconfig down instead of ifconfig detach for STREAMS-based X.25 in AIX Version 3. This does not apply to Version 4 users.

    Map IP Addresses to X.25 NUAs

    AIX Version 3.2 conforms to RFC 1236 "IP to X.121 Address Mapping for DDN" which defines an algorithm that automatically determines the NUA based on the Internet address.

    Attachment to the Defense Data Network (DDN)

    Choosing the DDN network identifier invokes this algorithm in place of the x25ip tables. The RFC defines this support for Class A Internet addresses only (Class B and C addresses are also supported).

    For All Other Network Identifiers

    For all other types of network, the x25ip table maps TCP/IP Internet addressing to the X.121 Network User Address (NUA). This table is maintained via the SMIT IP/X.25 Host Entry menu.

    Create an entry in this table for each remote host using either of the following methods:

    Both of these methods display the Add an X.25 "Licensed Program" Product IP Host Entry screen. This screen contains fields in which you type or select the desired values. The following figure shows the entry fields with sample field values:

    Notes:
    1. A virtual call is made when the first session with a remote system is established.
    2. Several IP sessions can share the same virtual circuit. Once an X.25 communication has been established, all IP traffic between two systems shares the same virtual circuit. Only one SVC or PVC is needed for communications between two systems.
    3. The X.25 virtual circuit is left connected for a period of time after the last IP session has been closed. This is so the cost of call establishment is not incurred if the remote system is contacted again in the near future (which is reasonable). This time delay, which is approximately 20 minutes, is modifiable with the Network Options (no) command using the arpt_killc option. The no command is available to modify some network parameters for the pending IPL. If you want your new values to remain unchanged after each system boot, add the command at the end of the /etc/rc.tcpip file. For example, to clear the virtual circuit after five minutes of inactivity:
      no -o arpt_killc=5
      Note: This applies to SVCs and PVCs that are removed using the arp -d command.
    4. To clear a TCP/IP X.25 virtual connection, use the arp -d command. For example to clear an existing virtual connection between kilix and filix , type on filix:
      arp -d filix

    Managing the Triple-X PAD

    The functionality of the PAD can be divided into two types of PAD, host and terminal PADs. The host PAD allows remote ASCII terminals to access applications running on it, with their access across an X.25 network. A terminal PAD allows the ASCII terminals to directly attach to the system over asynchronous links, and for them to use the X.28 protocol to establish calls with the remote X.25 attached system.

    To enable host PAD support, the Manage Triple-X PAD option should be selected from the Manage X.25 Device Driver menu. This option brings up a menu from which the host PAD support can be added or stopped - the Start/Stop the Triple-X PAD X.29 Daemon menu. Select the Add and Configure/Stop the X.29 Daemon menu to start the X.29 daemon. The daemon is only started for that instance and will be stopped automatically if the machine is rebooted. To avoid having to restart the daemon each time, add the following to any start-up script:

    /usr/lib/drivers/pse/x29d

    The terminal support does not need to be configured and is available once the X.25 and PAD software is installed, and the X.25 ports configured.


    [ Previous | Next | Contents | Search ]