This subsection explains the correspondence between the source HiRDB and the source Datareplicator.

(a) Proportion when source HiRDB is a single server

Source HiRDB (single server) to source Datareplicator = 1 to 1

There can be only one source Datareplicator per source HiRDB. There can be only one source HiRDB per source Datareplicator.

(b) Proportion when source HiRDB is a parallel server

Source HiRDB (parallel server) to source Datareplicators = 1 to n^#

#: n = 1 + number of servers containing a back-end server

For one source HiRDB, one source Datareplicator must operate at each server that contains the system manager or a back-end server. A back-end server need not contain a database that is subject to extraction processing.

On the other hand, each source Datareplicator can serve only one source HiRDB.

To establish correspondence between source and target Datareplicators, you must design the items discussed below.

(a) Proportion

Source Datareplicator to target Datareplicators = 1 to m

Source Datareplicators to target Datareplicator = n to 1

Variable	UNIX	Windows
m	1 to 64#	1 to 63#
n	1 to 128	1 to 63

#

If sendmst is specified in the sendcontrol operand, the value range is 1 to 4,096.

You can send update information from a single source Datareplicator to multiple target Datareplicators. If the HiRDB is a parallel server, you can send update information to multiple target Datareplicators per source Datareplicator at a server containing a back-end server.

On the other hand, a single target Datareplicator can receive update information from multiple source Datareplicators.

If you run multiple target Datareplicators for one target HiRDB, you can receive update information from 128 or more (in Windows, 63 or more) source systems (data linkage identifiers). However, the following problems might occur:

• Import processing might result in lock errors.
• The order in which the update information is imported might not be as expected.

(b) Target identifiers

The identifier used by the source Datareplicator to identify the destination of update information is called a target identifier. Specify all target identifiers in the sendidxx or sendidxxxx operand in the extraction system definition and specify a target identifier for transmission of specific update information in the transmission statement in the extraction definition.

You can identify the transmission target by creating a transmission environment definition for each target identifier and specifying the target identifier in the sendhdsid operand in the transmission environment definition. The source Datareplicator can start transmission processing or specify reduced-mode operation for each target identifier.

You must specify the sendidxx or sendidxxxx operands consecutively in ascending order. To add a sendidxx or sendidxxxx operand during operation, you must initialize the source and target environments. To provide for the possibility of a future increase in the number of destinations, specify absent numbers (by specifying ** as the target identifier for each absent number) with the sendidxx or sendidxxxx operand. When you provide absent numbers, you can establish data linkage with a specified target simply by initializing the target (there is no need to initialize the entire environment at the source and target).

(c) Data linkage identifiers

The identifier used to establish correspondence between source and target systems for data linkage is called a data linkage identifier. Specify the same data linkage identifier in the source and target systems between which data is to be linked. A specified data linkage identifier must be unique in the target system. For a source Datareplicator, use the dsid operand in the extraction environment definition; for a target Datareplicator, use the dsidxxx operand in the import system definition.

If the source system is HiRDB, specify a data linkage identifier for each extraction unit. An extraction unit is a unit of update information extraction and transmission. When the source HiRDB is a single server, a single source Datareplicator achieves extraction and transmission; in this case, one source Datareplicator is equivalent to one extraction unit. On the other hand, when the source HiRDB is a parallel server, the source Datareplicator for each back-end server achieves extraction and transmission; in this case, the source Datareplicator for each back-end server is equivalent to one extraction unit.

You must specify the dsidxxx operands consecutively in ascending order. To add a dsidxxx operand during operation, you must initialize the source and target environments. To provide for the possibility of a future increase in the number of destinations, specify absent numbers (by specifying ** as the data linkage identifier for each absent number) with the dsidxxx operand. When you provide absent numbers, you can establish data linkage with a specified source simply by initializing the environment of the source to be added (there is no need to initialize the entire environment at the source and target).

(d) Source Datareplicator identifiers

The identifier used to identify a source Datareplicator is called a source Datareplicator identifier. Specify the source Datareplicator identifier in the hdeid operand in the extraction system definition.

(e) Target Datareplicator identifiers

The identifier used to identify a target Datareplicator is called a target Datareplicator identifier. Specify the target Datareplicator identifier in the hdsid operand in the import system definition.

At the source Datareplicator, use the sendhdsid operand in the transmission environment definition to specify the target Datareplicator identifier at the destination of update information.

(f) Service name

To transfer update information between source and target Datareplicators, you must add the name of the communications service in the following services file:

• UNIX Datareplicator: /etc/services
• Windows Datareplicator: Windows-system-directory\drivers\etc\services

Specify the name of the applicable communications service in the hdeservices operand in the transmission environment definition for the source Datareplicator.

(g) Host name

To send update information, you must register the host name in the following hosts file at the source system:

• UNIX Datareplicator: /etc/hosts
• Windows Datareplicator: Windows-system-directory\drivers\etc\hosts

Specify the host name in the hdehost operand in the transmission environment definition.

To establish correspondence between target Datareplicator and target HiRDB, you must design the items discussed below.

(a) Proportion

Target Datareplicators to target HiRDB = n to 1

There can be only one target HiRDB for each target Datareplicator. On the other hand, there can be multiple target Datareplicators per target HiRDB.

(b) HiRDB authorization identifier and password

A target Datareplicator issues SQL statements and executes import processing in the form of a UAP that runs on the target HiRDB. Therefore, a HiRDB authorization identifier and password are required in order to establish connection with a target Datareplicator. To specify a HiRDB authorization identifier and password, use the hirdbusr operand in the import system definition. For details about the import system definition, see 5.8 Import system definition.

You can also enter the password from the standard input when the target Datareplicator is started.

The following figure shows an example definition of identifiers when data is linked from one source Datareplicator to two target Datareplicators.

Figure 4-17 Example definition of identifiers

Explanation:

• This example specifies the source Datareplicator identifier and the target Datareplicator identifier in the hdeid operand in the extraction system definition and in the hdsid operand in the import system definition, respectively.
• The example specifies the data linkage identifier in the dsid operand in the extraction environment definition and in the hdsid operand in the import system definition. Specify the same data linkage identifier for both source and target systems that link data.
• The example specifies the target identifier in the sendidxx operand in the extraction system definition. Create a transmission environment definition for each target identifier, and then specify the target system identifier (target Datareplicator identifier) in the sendhdsid operand.