UPIC-R uses transport systems for communication. In almost all practical situations, this involves TCP/IP with the protocol referred to as RFC1006. Transport systems have their own address regulations. The RFC1006 protocol is characterized by the fact that each transport system application signs itself on to the transport system with a name, referred to as the transport selector (T-SEL). The partners address one another using these names. RFC1006 is based on TCP/IP, so TCP/IP also requires the following addressing information:
System name
Port number
There is no general recommendation with respect to the port number on Unix, Linux and Windows systems. Port number 102 should, however, be used with care.
UPIC-R is configured using local_name and partner_LU_name, with the local_name being mapped on the local T-SEL. The application name from the two-part partner_LU_name is mapped on the remote T-SEL, the system name from the two-part partner_LU_name is the name of the system in the network. The partner_LU_name must be two-part, otherwise the described procedure does not work.
When mapping the local_name and the application name to the T-SEL, bear in mind that the character code of the T-SEL is not defined a priori. The two systems on which the server and client are running can use different character codes for representing the T-SEL (e.g. Windows systems uses an extended ASCII character code, BS2000 systems the EBCDIC character code). Consequently, the format of the names must be defined. Three character formats are possible between UPIC and UTM: ASCII, EBCDIC and TRANSDATA. The TRANSDATA character set is a restricted subset of the EBCDIC character set. UPIC-R checks whether the character set used by local_name and/or the character set used by the application name can be converted into the TRANSDATA character set. The TRANSDATA character format is used if this is the case, otherwise the EBCDIC character format is used.
One port number each is assigned to both the local_name and the partner_LU_name. The two port numbers are not derived from the name, they are always set to the value 102 by default.
The local port number is assigned to the local_name. The default value can be overwritten. The local port number is a purely formal value which does not have any effect, and is only entered on grounds of compatibility. It should be disregarded in the configuration of UPIC-R.
The remote port number is assigned to partner_LU_name. In contrast to the local port number, there is a significant importance attached to the remote port number. This is because the UTM partner application is addressed using the remote port number.
BS2000 only
In the vast majority of practical cases, it is sufficient to use the default value 102. BCAM and CMX always support port 102 as the central access port for RFC1006. Although it is possible to select another port, this requires a significant amount of configuration work on the server side, for example BCMAP entries have to be created for the BS2000 system. Such configurations require a certain level of experience and are not described here. As a rule, port 102 cannot be used if the UTM partner application is running on a system which uses PCMX to access the transport system. In this case, the value of the remote port number must be overwritten with the value which is used by the UTM application.
The values T-SEL, T-SEL format and local port number of the local_name can be overwritten with the following calls:
Specify_Local_Tsel
Specify_Local_Tsel_Format and
Specify_Local_Port
The values can also be overwritten by entries in the upicfile. In this case, the corresponding values are defined using keywords. The upicfile is described in section “The side information file (upicfile)”.
The addressing information for the network can be formed by specifying the local_name and using the internal rules of UPIC to have the network address created. It is also permitted and a function has been provided to overwrite one or more of the values derived from the local_name using the specified calls. It is permitted for any mixture of derived, default and explicitly set values to be used in this case. Equally, it is permitted for all of the values derived from the local_name to be overwritten. The local_name is meaningless if you select this type of configuration. You can then specify any local_name whatsoever, only providing it is compliant with the formal criteria of the Enable_UTM_UPIC call.
The values system name (or the Internet address derived from it), T-SEL, T-SEL format and remote port number can be overwritten with the following calls:
Set_Partner_Host_Name
Set_Partner_IP_Address
Set_Partner_Tsel
Set_Partner_Tsel_Format
Set_Partner_Port
The Set_Partner_Host_Name call is ignored if the Set_Partner_Host_Name and Set_Partner_IP_Address calls are both called. The values can also be overwritten by entries in the upicfile. In this case, the corresponding values are defined using keywords. The upicfile is described in section “The side information file (upicfile)”.
In many cases, the addressing information for the network can be formed by specifying the partner_LU_name and using the internal rules of UPIC to have the network address created. It is also permitted and a function has been provided to overwrite one or more of the values derived from the partner_LU_name using the specified calls. It is permitted for any mixture of derived, default and explicitly set values to be used in this case. Equally, it is permitted for all of the values derived from the partner_LU_name to be overwritten. The partner_LU_name is meaningless if you select this type of configuration. You can then specify any partner_LU_name whatsoever, only providing it is compliant with the formal criteria which are required of it (among other aspects, it must be two-part).