View previous topic :: View next topic
|
Author |
Message |
rocky_balboa
New User
Joined: 24 Mar 2010 Posts: 61 Location: Cape of Good Hope
|
|
|
|
Hello
We have a requirement to migrate customer data from one of the systems( which is going to sunset. lets call it 'A') to another system which has IMS/DB( call it 'B'). B has transcations which get triggered by web intereface which add/update customer information. Web - > MQ - > IMS Queue -> IMS Tran -> DB
Now possible solutions for migrating the data :
1) Write a batch program having the same rules as those of transactions. This will require analysis of all the rules how each and every segment gets modified and test all of it.
2) Trigger the IMS transactions feeding them input in a format which they require through batch.
Problems here - Where to write the transactions statistics - customers,addresses updated deleted etc. Also , if a customer update fails it will be a huge task to analyse production dumps.
I prefer the second one since it reduces test effort dramatically and possible errors in customer data migration and most importantly the volume is in a few thousands. However, since I am new to IMS ( have been a CICS DB2 guy) I am not sure what best options are there for the second solution.
Any pointers/suggestions for the approach or considerations which I might have missed will be really helpful.
Thanks |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
Quote: |
We have a requirement to migrate customer data from one of the systems( which is going to sunset. lets call it 'A') to another system which has IMS/DB( call it 'B'). B has transcations which get triggered by web intereface which add/update customer information. Web - > MQ - > IMS Queue -> IMS Tran -> DB |
it is not clear what customer data You are talking about ...
if a transaction, whatever the logic path, adds/updates customer data in an IMS database
it means that the data is already there , so no need to migrate anything.
unless You clarify , the chances of getting help are pretty slim! |
|
Back to top |
|
|
rocky_balboa
New User
Joined: 24 Mar 2010 Posts: 61 Location: Cape of Good Hope
|
|
|
|
My bad that it caused confusion.
The data to be migrated is of a different application having its own customer data in their own DB.( Referred to it as 'A' in the first post). This data has to be migrated to another application which runs on IMS DB.(This is 'B').
When I mentioned That "B has transcations which get triggered by web intereface which add/update customer information. Web - > MQ - > IMS Queue -> IMS Tran -> DB " I just gave an idea how transactions on B run. These transactions do not propogate data from application A to B.
A and B are not connected to each other in any way. |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10873 Location: italy
|
|
|
|
OK, let' s see ...
system A has some application which deals with some customer data,
( inquiry,updates,add,delete)
system B has some application which deals with some customer data
(inquiry,add,delete,update)
why cannot the migrated data be updated using the same logic of the existing application on B ?
Why not just write a run once application to convert the data from the old(A) structure/layout to the new(B) structure
and use the existing b application to update the data ??? |
|
Back to top |
|
|
rocky_balboa
New User
Joined: 24 Mar 2010 Posts: 61 Location: Cape of Good Hope
|
|
|
|
They will be migrated using the application logic in B. The logic for customers' maintenance is in the IMS transactions which apart from customer maintenance, do other things as well.
Now possible solutions for migrating the data :
1) Write a batch program having the same rules as those of transactions. This will require analysis of all the rules how each and every segment gets modified and test all of it.
2) Trigger the IMS transactions from batch feeding them input in a format which they require.
Problems here - Where to write the transactions statistics - customers,addresses updated deleted etc. Also , if a customer update fails it will be a huge task to analyse production dumps.
I prefer the second one since it reduces test effort dramatically and possible errors in customer data migration and most importantly the volume is in a few thousands. However, since I am new to IMS ( have been a CICS DB2 guy) I am not sure what best options are there for the second solution.
Any pointers/suggestions for the approach or considerations which I might have missed will be really helpful.
Thanks |
|
Back to top |
|
|
rocky_balboa
New User
Joined: 24 Mar 2010 Posts: 61 Location: Cape of Good Hope
|
|
|
|
Is my post so confusing still? |
|
Back to top |
|
|
Rohit Umarjikar
Global Moderator
Joined: 21 Sep 2010 Posts: 3053 Location: NYC,USA
|
|
|
|
second one is better to make it. write an error file out if the trx fails to troubleshoot the issue. |
|
Back to top |
|
|
|