Virtual Data Marts

by Antonio M
This author talks about Data marts and how they are required for optimizing data
accessibility. He describes a method to create data marts for the use of
local decision support software without the data in the data mart having to be
stored locally. By doing this it can save storage space and data transmission
resources.

When you have data stored in data warehouse and then that data gets stored in data marts
there is duplication of data. Not only that but if data in a data warehouse gets
updated then data in a data mart should also be update as well if there was any data
taken from a data warehouse to begin with. Something like this is very redundant and
can allow for inconsistent data. In order to solve something like this there would need
to be a function where if one location is updated, say for example a Data Mart,
the data warehouse should also be updated so that there is consistent data and the same
should go if the data warehouse were to be updated then the Data Mart should automatically
be updated as well. Instead of doing all these updates a solution that the author
proposes is to have a Virtual Data Mart.  Regular Data Marts use local storage resources
to just copy data from a data warehouse into the Data Mart. A Virtual Data Mart would instead
use a metadata table structure that would then reference data elements within the data warehouse.
The name of this metadata table is called a “Metapoiner Table” . This is where metadata
that is associated with data elements is recorded. The benefit of this is the data element
that is wanted is just stored in one location now and there is also no local disk storage needed
for the data mart. The meta pointer table resides within the data warehouse which is also where
the target data mart resides.

I fond this article to be very interesting because it talks about another way we can optimize
the way data is managed. I think the author made some very interesting points and examples on
how a data mart can be virtual in order to save space.

I think this relates to our class because it talks about data marts and data warehouses and
the role that they play in the use of data through a decision support system. I think that
this article would be good to read for students like us learning about data bases and especially
those students that want to get a career in databases. This article was very helpful and gave
a lot of ideas to me personally on how to possibly make a Virtual Data Mart.

Reference:
Hodge, L.K.; Milligan, M.; , “Managing virtual data marts with metapointer tables,” System
Sciences, 2003. Proceedings of the 36th Annual Hawaii International Conference on , vol., no.,
pp. 7 pp., 6-9 Jan. 2003 doi: 10.1109/HICSS.2003.1174606 URL: http://0-ieeexplore.ieee.org.opac.
library.csupomona.edu/stamp/stamp.jsp?tp=&arnumber=1174606&isnumber=26341

3 thoughts on “Virtual Data Marts”

  1. Nice article, very informative! I know duplicate data are bad for the database, I wastes space and also decrease performance of the database. I am glad someone is coming up with the idea of virtual data marts, it seems like a great solution to the problem with duplicate data, but keeping a large amount of data in one location does make me wonder what if something happens to that location like earthquake, or any natural disaster, then we will be disconnected from the data that is located there.

  2. I really liked your article, I can only imagine how redundant it must be to update and delete entries from the data marts, using the old method. Not only is the virtual data mart much more convenient but it also much economic. Having to store all that extra data into physical space can be very expensive. It also saves time on the admin to focus his attention to other aspects of the database.

  3. this article definitely was interesting. since there is so much data, it would be very tedious to update information. not to mention if data duplication exists, it would cause a big headache when we find out we didn't update all the data. it sounds like data marts would be very helpful to manage data and time.

Comments are closed.