Data Mart|Maintenance of data mart

Maintenance of data mart

Data Mart provide complete information about a particular subject to the users. The users can find solution to all their queries related to that particular subject. Hence, the proper & adequate, sufficient maintenance of a data mart id necessary.

Periodic maintenance of a data mart means loading, refreshing that data in it. Refreshing may be done on daily, weekly, monthly, quarterly or yearly basis. Following are the points that can help in proper maintenance of data mart.

 1)Upgrading of data mart from data sources:  The data mart should be upgraded at regular intervals from operational data sources. So, that all the latest data related to the particular object can be retrieved easily & the users can efficiently solve their problems. Hence the data mart provides the adequate knowledge to the users & is maintained properly. The data can not only be taken from operational data source but it can also be taken from data warehouse.

  2)Backup Management: The backups are taken for only one reason:

Protection against data loses: Backups are useless if they can’t be used when they are required backup package have their own tools for managing backup & recovery.

Features:

Scheduling

Backup data tracking 

Data  Base awareness

 3) Metadata Management:  Mata data is data about data. It is like a card index describing how information is structured within the data warehoused. Metadata is used for a variety of purpose. The structure of meta data will different between each process because the focus is different. Hence the meta data should ne managed properly & efficiently so that the data mart produce accurate result. 

  4)  Hardware and Software requirements: Data mart  require additional hardware & software that are used to run different applications. The data mart database will require additional processing power & storage capacity to handle the user queries. Data mart should not be used as an alternative to aggregation. This will increase the costs hence the hardware & software should be maintained in such a way that overheads can be easily avoided.

  5) Network access: The process of loading data into each data mart will be affected by the available physical connections between both items of hardware. In many cases each data  mart will be in a different geographical location from the data warehouse.

Hence, such issue should be taken care of & should be handled in an efficient manner So, the data mart are not defected

 6)Time window constraints: The determination of how many data mart are feasible will vary, depending on a number of factors namely

• Time window available
• Network capacity
• Volume of data being transferred
• Mechanism being used to insert

We should calculate the approximate time necessary to complete the data mart load in order to determine . if the architecture is feasible

In time window constraints we have to:

1)calculate the data volume being sent into each data mart

2)Calculate elapsed time to transfer access the LAN & WAN into each mart

3)Calculate elapsed time to load the data into each mart