DATA MANAGEMENT
…covering everything which includes the word "data" or "information"
MDM Data Hub Styles: Repository Style Hub – Part 1
5 Sep 2011
In a repository style data hub all the data is copied from the source systems to the a single data hub and at the same time duplicates are removed through matching and linking. This data hub style is also known as a transactional data hub.
This means that all systems need to be modified to use the new data hub. In most cases, this tends to be a very complex, time-consuming and costly exercise. The more systems and the more complexities these systems have, the more difficult, time-consuming and costly the task. If you have a system which comes with a built-in database or which has been built using non-standard method such as proprietary programming languages then this will be a particularly challenging task and you may need to seek help from the system vendor. Furthermore, when building the single data hub all data sources have to agree on a common data model which can be both technically but more often politically challenging as each source system wants their data model to be supported by the data hub.
On the other hand, the advantages are that you will have a single, high data quality and reliable data source which is accessed by all systems. There will also not be any synchronisation or latency issues. Plus a single, central data hub is easier to manage than several, dispersed data sources.
Although the advantages are attractive, it is often very difficult to get approval from the business sponsors to go ahead with this approach. Even if you do get approval, it is a very time-consuming approach to implement an MDM solution and you may risk proving quick ROI to the business.
No related posts.
Comment Form
About this blog
The purpose of this blog is to help you learn/expand on your current knowledge about the various domains which are part of Data Management. This includes Business Intelligence, Master Data Management, Data Quality, Data Governance, Data Integration etc.
About the author
I am Manjeet Singh Sawhney and work as an Information Management Consultant for Tata Consultancy Services (TCS) in London (UK). Prior to TCS, I have worked for Accenture, Tibco Software and Initiate Systems (now IBM). My areas of expertise are Master Data Management (Customer, Product, Reference), Metadata Management, Data Governance, Data Quality, Data Integration, Data Profiling and Data Analysis. I am using this blog to share my knowledge and hope that you will find it useful. Thanks for visiting.
Advertisement & Sponsorship
If you wish to advertise (banner or text links) on this blog or sponsor any blog posts then contact the author using the 'Contact the Author' form below. Please include some details about your advertisement / sponsorship request and a response will be sent to you within 24 hours.
- Challenges in MDM | DATA MANAGEMENT: [...] has grown. So I thought of writing this post which will build on the already published post; C [...]
- Manmohan: Informative article on how business intelligence is used to analyse customer data to find suitable c [...]
- Manjeet Singh Sawhney: Thanks very much and I am glad I could help and you that you found the articles useful. Wish you [...]
- belomomo: Hi Manjeet Singh Sawhney, Thanks for your reply and help. Actually, I read some books about MDM es [...]
- Manjeet Singh Sawhney: Thanks very much for your feedback. I have published the last article of the series at http://dat [...]
- Challenges in MDM
- The Common Implementation Approach found in CDI Implementations
- How CDI can benefit Organisations – Part 4
- How CDI can benefit Organisations – Part 3
- How CDI can benefit Organisations – Part 2
- How CDI can benefit Organisations – Part 1
- Why Customer Data Management is so challenging
- What is Data Profiling?
- Process for implementing MDM within an Organisation – Part 3
- Process for implementing MDM within an Organisation – Part 2
3 Responses to MDM Data Hub Styles: Repository Style Hub – Part 1
diablo 3
September 6th, 2011 at 6:48 pm
Everything is really open and quite clear explanation of concerns. was truly data. Your web site is really useful. Many thanks for sharing.
Girish Nair
September 11th, 2011 at 8:57 pm
Hello Manjeet,
Good article on Hub style. This type of MDM implementation can also be termed as Consolidation style of MDM.
I have a question on the cons of this style of implementation. Why do you think this type of implementation is complex and time consuming.
i think this would be less complex as the source system do not need to change anything inside their system. They just need to provide data to MDM using a good ETL support. MDM needs to work harder to understand it and apply to the model which will support all other systems.
Regards
Girish
Manjeet Singh Sawhney
September 17th, 2011 at 12:17 pm
Hi Girish,
Thanks for your feedback. The reason why I believe a repository style data hub is more complex and time consuming is because of several reasons.
In order to create a repository style state data hub, all the data must be integrated from several source systems into a single data hub. These source systems and data may have different data models and therefore when a single data model for the data hub must be created then it will be time-consuming nad challenging to agree with all sources to a single, common data model.
Similarly, if an organisation has purchased a product which comes with a built-in database then “detaching” it from the system and change the system such that all data references are made to the MDM data hub can both be time-consuming and complex. The organisation may need to seek help from the system vendor to assist with this task. This can also be quite costly that may not have been taken into account when valuing the cost of the MDM implementation.
Building a repository style hub will certainly take more time than any of the other data hub styles, because all data needs to be moved into the single data hub. Therefore there will be more – perhaps sophisticated – ETL rules compared to the other two data hubs – registry and hybrid – where all/some data is kept in the source systems, respectively.
Lastly, systems where the data is moved to the data hub must be updated to point to the data hub as their new data source. Updating all system will certainly be a very time-consuming and labour intensive task.
I hope I have made a good effort to try and answer your question. Please let me know if you have any further questions related to this or any other article which I have published on my blog.
Thanks,
Manjeet