File routing and file combining are data management techniques supported by MIMIX user journal replication processes in data groups that use single-threaded database apply processing. Because both techniques require keyed replication, they are not possible in data groups that use multithreaded database apply processing.
The way in which data is used can affect the configuration requirements for a file routing or file combining operation. Evaluate the needs for each pair of systems (source and target) separately. Consider the following:
-
Does the data need to be updated in both directions between the systems? If you need bi-directional data flow, see topic Configuring for bi-directional flow.
-
Will users update the data from only one or both systems? If users can update data from both systems, you need to prevent the original data from being returned to its original source system (recursion).
-
Is the file routing or file combining scenario a complete solution or is it part of a larger solution? Your complete solution may be a combination of multiple data management and data distribution techniques. Evaluate the requirements for each technique separately for a pair of systems (source and target). Each technique that you need to implement may have different configuration requirements.
File combining is a scenario in which all or partial information from files on multiple systems can be sent to and combined in a single file on a target system. In its user journal replication processes, MIMIX implements file combining between multiple source systems and a target system that are defined to the same MIMIX installation. MIMIX determines what data from the multiple source files is sent to the target system based on the contents of a journal transaction. An example of file combining is when many locations within an enterprise update a local file and the updates from all local files are sent to one location to update a composite file. The example in Figure 16 shows file combining from multiple source systems onto a composite file on the management system.
Figure 16. Example of file combining
To enable file combining between two systems, MIMIX user journal replication must be configured as follows:
-
Configure the data group definition for keyed replication. See topic Keyed replication.
-
If only part of the information from the source system is to be sent to the target system, you need an exit program to filter out transactions that should not be sent to the target system.
-
If you allow the data group to be switched (by specifying *YES for Allow to be switched (ALWSWT) parameter) and a switch occurs, the file combining operation effectively becomes a file routing operation. To ensure that the data group will perform file combining operations after a switch, you need an exit program that allows the appropriate transactions to be processed regardless of which system is acting as the source for replication.
-
After the combining operating is complete, if the combined data will be replicated or distributed again, you need to prevent it from returning to the system on which it originated.
File routing is a scenario in which information from a single file can be split and sent to files on multiple target systems. In user journal replication processes, MIMIX implements file routing between a source system and multiple target systems that are defined to the same MIMIX installation. To enable file routing, MIMIX calls a user exit program that makes the file routing decision. The user exit program determines what data from the source file is sent to each of the target systems based on the contents of a journal transaction. An example of file routing is when one location within an enterprise performs updates to a file for all other locations, but only updated information relevant to a location is sent back to that location. The example in Figure 17 shows the management system routing only the information relevant to each network system to that system.
Figure 17. Example of file routing
To enable file routing, MIMIX user journal replication processes must be configured as follows:
-
Configure the data group definition for keyed replication. See topic Keyed replication.
-
The data group definition must call an exit program that filters transactions so that only those transactions which are relevant to the target system are sent to it.
-
If you allow the data group to be switched (by specifying *YES for Allow to be switched (ALWSWT) parameter) and a switch occurs, the file routing operation effectively becomes a file combining operation. To ensure that the data group will perform file routing operations after a switch, you need an exit program that allows the appropriate transactions to be processed regardless of which system is acting as the source for replication.