This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

DEVELOPERS FORUM - ARAS in an online/offline scenario? master data management

fli - Tuesday, March 29, 2011 4:40 AM:

Dear members of the forum,
Is it possible to use ARAS in an online/offline scenario?
The case is that I want to take a version of the software with me on a laptop without being connected to any network, enter new data or alter existing (CRUD) and be able to syncronise those data with the “online” data when I get back to the internal network.
Also I would like to know if ARAS supports some kind of master data management so that if the data model has changed while being offline, these changes will be accommodated for when syncronising back.
BR/ Christoffer


PeterSchroer - Tuesday, March 29, 2011 8:25 AM:

Hi Christoffer,

I have made an implementation for off-line data access/editing.    In this case,  I run an AML query to extract a large dataset,  store this XML file on the C: drive of the user.   Then we developed a stand-a-lone Silverlight (Windows WPF) application that reads this XML structure,  and allows viewing and updating.   When the User is connected again to the server,  he runs a Synchronize action,  and the XML is sent back to a Server method,  that (1) looks for new data which is added to the master and then (2) looks for modified data ,  which must be merged, looking at date stamps etc. with the master data.    

Your idea to use Innovator also on the off-line machine is also good.   This is the same as running 2 or more Innovator instances for replicated/distributed data management.     Synchronizing the datamodel is the easiest step.     IF  (very important)  you have good discipline of using Packages for all your development.   When you want to re-syncronize the data model,  then just Export-Import the packages again.       Data can also be  Export / Import but you have to think carefully what your rules are.   How do you handle conflicts?  Data has been edited by 2 different people in 2 different Instances.      Not impossible,  but you need to have a careful design,  and good agreement from the end-users how these conficts will be resolvedd.  Moving data, and updating data is the easy part.   Resolving conflicts requires everyone agrees how they might lose edits.

One customer (who was not using Packages for development)  used a SQL nightly transfer to move the master data model,  and also useful things like User, Identity, List items.   This works as well as the Package Export/Import,  but the packaging tools make a simple job of loading data in the right dependency order.  If you work with SQL directly,  you need to think carefully about the sequence of steps  so that dependcies are created in the correct order.

Good luck - and please post the results !

Peter



fli - Thursday, March 31, 2011 3:23 AM:

 

Hi Peter,

Thank you for the answer.

It seems like a good project.

Christoffer