Hi,
Presently I have script #!/bin/ksh in which 4 executables are executed
as 1,2,3 and 4.Executable 2 is an extract program which extracts records from a table.and executable 4 is a program that updates a database.
Currenlty this process takes a hell lot of time and my aim is to reduce the overall execution time.
I want to change executable 2 to process only say 100 records.Suppose there 500 records in a database,I want to run 5 instances of executable 2 in parallel.I also want to make use of these 5 instances to update my database by passing the output of each of the 5 instances to executable 4.In other words,I also want to split executable 4 into same number of instances as executable 2.

I have another problem here :
I want to merge the output files of 5 sub program ,attach a header and trailer and then FTP it.Is this going to be time consuming ?

Can anybody help me how it is possible ? Is this approach correct to cut down the timing of the entire process ?
Please advise.
Thanks
Unixcurious

Link removed

We do this all the time for database batch processing.
The executables are going to have to be able to take parameters - from the command line, a file, or an environment variable. Then the parameter is passed into the executables as they run.

For example -

When we bill, the billing cycles are broken into 12 sub-cycles. This makes the process run ~12 times faster - if there are 12 cpu's and the database supports high volume transaction rates - we use Oracle. Our cycle bill time is 30 minutes instead of six hours.
pseudocode example:
When all the jobs are done we ftp the files over to another box to print bills.

tushar_johri, please don't post your webite link in a psuedo signature like that. You can post a link to your in our News, Links, Events, and Announcements forum if you want.