Would you pl thro some lights on the following w.r.to process chain.pl do not forward any link .
2.difference between connection type and process
3.collection process and how to do them in system?
See this ,
Q . would like to know how I can start a process chain in a specific program without using an event ?
A. can use function module RSPC_CHAIN_START or RSPC_CHAIN_EXECUTE_REMOTE.
2) note 534630-Parallel processing of Change Run .
Process Chain Details :
To reduce the runtime of the change run, you can distribute adjustment of the aggregates across several work processes running in parallel. In this case, a separate work process (in which all aggregates belonging to the InfoCube are adjusted) is started for each InfoCube whose aggregates are affected by the change run. These additional work processes are interactive work processes.
When implementing this solution, bear in mind that the runtime of the paralleled version of the hierarchy and attribute realignment run depends on the adjustment times of aggregates on the different InfoCubes. If the adjustment times of the aggregates are distributed evenly across the different InfoCubes, you can significantly improve performance with parallel processing. However, if one InfoCube and its aggregates occupy most of the runtime of the conventional change run, the speed gain during the parallel change run declines accordingly.
If you do not make any changes, the change run continues to be executed serially in a background process. To use the paralleled variant of the change run, you must refer to or carry out the following points:
(Optional) Create an RFC server group in transaction RZ12. For detailed instructions on using RFC groups, see the SAP Library. (Before you can use RFC groups, at least one operation mode must be created and active for normal operation in transaction RZ04.)
When creating the RFC group, you must make absolutely sure that the name is fully entered in upper case and that no special characters or spaces are used.
After the RFC server group is defined, you must use the maintenance report SAP_RSADMIN_MAINTAIN to create the object entries CR_RFCGROUP with the name of the corresponding RFC server group and CR_MAXWPC with the maiximum number of work processes for use in the table RSADMIN. The CR_RFCGROUP entry is optional. Run the report SAP_RSADMIN_MAINTAIN (transaction SE38) with the two parameters, object = "CR_RFCGROUP" and Value = <Name of the RFC- server group> (optional) or Object = "CR_MAXWPC" and Value = <Number of work processes>.
The number of the work processes used depends on the number of InfoCubes with aggregates that have to be adapted. The same number of work processes is started as the number of InfoCubes affected by the change run. You can restrict this number with the object entry CR_MAXWPC. (This may apply, for example, if there are too many InfoCubes with aggregates). Excessive parallel processing may cause problems in the memory consumption and cause large-scale paging.
In defining the RFC server group, you can control the value distribution of the work processes across several instances of the BW system. You can define the maximum number of work processes to be occupied per application server as a percentage on the RFC server group. Alternatively, you can also define the number of interactive processes that you want to remain free. You should select these parameters very carefully to prevent the system from becoming unusable as a result of many busy interactive work processes. Remember that the settings that you make are valid for all RFC calls and not just for the change run.
If the CR_RFCGROUP parameter is initial, all application servers are used.
Since the additional work processes are interactive work processes, you may also have to adapt the profile parameter rdisp/max_wprun_time for the affected application server to prevent the work processes from terminating if the timeout limit is reached prematurely. (A work process is active until all aggregates of an InfoCube are adjusted.)
Furthermore, the temporary tablespace of the database may have to be increased because of parallel processing. If aggregates are set up during the change run, you can also limit the size of the blocks that are read at the same time (In Release 3.X: Transaction SPRO > SAP Reference IMG > BW Customizing Implementation Guide > Business Information Warehouse > General BW Settings > Parameters for aggregates > Block size parameters or in Release 2.X: Report RSDDK_BLOCKSIZE_SET). You can thus reduce the consumption of temporary tablespace.
For more details on the size of temporary tablespaces on Oracle, see note 359835.
You should also note that parallel processing of the change run increases the memory consumption of the SAP server. A paralleled solution for the change run is not recommended if there is already a CPU or memory bottleneck on the DB server. A fast I/O subsystem is required.
(For Oracle only) Note that an application server process may use several database processes (parallel query slaves). To prevent the first application server from occupying all database processes, you must either ensure that there are enough database processes or you must restrict the number of database processes per application server process (see note 544521)
Hope It HElps....Reward points if its useful.
A Process is characterized by
Process Type Kind of Process. The process type is set in the RSPROCESSTYPES view.
Process Variants It is the name of the process. It displays the configuration of a process of a particular type set at the time of definition. A process can have various variants.
Process Instance - The process instance is the characteristic value of the process It contains the most important information that the process, or subsequent processes wants to communicate. The logs for the process are stored under the process instance
Collection Process - Allows several chain strings to be combined to form one individual string.
A collection process collects several chain strings to form one string in the process chain maintenance. The following collection processes are available in the process chain maintenance:
And Process (Last) - This process does not start before all events of the predecessor processes, that is including the last event, that it has waited for, have been successfully triggered.
Or Process (Every) - The application process starts every time a predecessor process event has been successfully triggered.
XOR Process (First) - The application process starts when the first event in one of the predecessor processes has been successfully triggered.