topic Re: MDM or Data Integration: Deciding factors in Talend Data Catalog

MDM or Data Integration: Deciding factors

Anonymous — Thu, 31 Mar 2016 02:15:12 GMT

Hi there,
I am just looking up Talend in general. I was wondering what would be the deciding factors between me starting to use MDM or the Data Integration (Open Studio).
This is an exploration at the moment - I do have a project on which to work, however it is difficult for me at this point to know the full scope of use for Talend. From what I have read so far, the DI is amply sufficient for my current use. However, am I right if I say that MDM has a larger scope, and that on the long run I might be better off starting with MDM? Are there jobs that only DI will perform, or am I right in seeing DI as kind of a subset of MDM? If that is the case, I think I would probably start with MDM right away.
ANy thoughts/advices on this very welcomed,
Franck

Re: MDM or Data Integration: Deciding factors

Anonymous — Tue, 05 Apr 2016 04:30:47 GMT

Hi,

Actually, Talend open studio for MDM product includes both TOS DI and TOS MDM bundles.

The DI integration perspective in Talend open studio MDM is as same as Talend Open studio Data Integration product.

Best regards

Sabrina

Re: MDM or Data Integration: Deciding factors

Anonymous — Thu, 07 Apr 2016 10:58:51 GMT

Hello Franck
From my humble point of view (as I was in the same situation as yours few months ago), here what I can tell you about what Talend should you use. As Sabrina told you Talend DI is integrated inside MDM. So you can start with Talend DI to create job, and then, if you think you would like to use MDM, reuse your workspace created inside DI in MDM studio. So don't worry about that point.
TALEND DI
In Talend DI you can design autonomous jobs (I deploy them as command line tool on unix server) that enable you to Extract and Transform data, and then Load it (in database, or file, or stdout, whatever you want).
The kind of easy transformation deals with row to row mapping (relational data).
If your data is an tree with recursives nested branchs of unknown deepness, then it is not so simple (I think there is a component in Talend Studo with subscription, but not sure of this point)

<item> Grand father item
  <item> Father Item
    <item> Child Item
    </item>
  </item>
</item>

TALEND MDM
Talend MDM is for a particular purpose : MDM. As such, you design a database schema, jobs that can load or extract data from that database, WebUI to access data on this database and (thats the point VERY different from DI) some "MDM server side jobs"
These jobs are triggered upon modification of the data in your master database that enable you to propagate the modification to other databases, or to validate the data before insertion. If you want to know about more about MDM ther is a VERY GOOD available (see the sticky topics of this forum). Take the time to do it, and you will know what MDM does. "Normal" jobs in TOS can adress more input/output kinds, but are not supposed (AFAIK) to be deployed on a server which purpose is to ensure coherency between data coming from multiple source.
Hope this help

Re: MDM or Data Integration: Deciding factors

Anonymous — Tue, 20 Sep 2016 20:58:42 GMT

Hey guys,
Awesome that helps a lot - I'm especially happy to know that I can probably for now stick with DI and migrate over to MDM if the need arises. Makes things simpler for now.
Thanks!

Re: MDM or Data Integration: Deciding factors

Anonymous — Thu, 22 Sep 2016 10:22:22 GMT

Hi there,
I am just looking up Talend in general. I was wondering what would be the deciding factors between me starting to use MDM or the Data Integration (Open Studio).
This is an exploration at the moment - I do have a project on which to work, however it is difficult for me at this point to know the full scope of use for Talend. From what I have read so far, the DI is amply sufficient for my current use. However, am I right if I say that MDM has a larger scope, and that on the long run I might be better off starting with MDM? Are there jobs that only DI will perform, or am I right in seeing DI as kind of a subset of MDM? If that is the case, I think I would probably start with MDM right away.
ANy thoughts/advices on this very welcomed,
Franck

Hi Franck,
of course not knowing your use case this is kind of hard to answer.
IMHO,at the lowest level a good mdm solution needs 5-6 things.
1) a way to integrate to sources
2) different way of loading that data ( webservices,batch,UI)
3) DQ and Match engines to check for duplicate or implement other usecases like blacklisting,households,etc
4) Security concepts to define who can change, modify,create what
5) a way to share the date from you MDM (webservices,batch,UI)
6) DQ to continuisly check the quality of your master data
Now.. you what you have to decide is ..if you want to build every yourself or buy a product that offers you these things..for instance Talend MDM.