implementation considerations

1. R.Preethi
M.sc(cs)
Nadar Saraswathi College Of Arts & Science.
Data Mining
Implementations considerations

2. Data Placements strategies
 Data warehouse grows, there are at least two options
for data placement.
 One is to put some of the data in the data warehouse
into another storage media ; e.g. , WORM , RAID , or
photo-optical technology.
 Second option is to distributed the data in the data
warehouse across multiple servers.

3.  The data selected for transport to the alternative storage
media is detailed and older and there are is less demand
for it.
 The bulk of storage can be handled either by the data
warehouse server or another server used for handling the
bulk of storage media.
 This configuration requires both corporation metadata
and the metadata managed for any given server.

4. Data Replication.
 Replication technology creates copies of databases on a
periodical basis, so that data entry and data analysis can
be performed are performed separately.
 Most companies uses data replications to copy the
important data to a separate data base replication server
to access it.
 The process can be used to move data into distributed
data marts.
 In this data replications OLAP tools used .

5. Database Gateways
 The traditional gateway technology provides LAN users
with the ability to easily access small amounts of
mainframe data .
 It is not optimized for moving large files.
 Networks can be slowed by multiple concurrent user
request for similar data through a gateway.
 Gateway access for decision support will often compete
with production applications for resources.

6. Metadata
 Metadata needs to be collected as the warehouse is
designed and built.
 The key to providing user and applications with a
roadmap to the information stored in the warehouse in
the metadata.
 It can define all data elements and their attributes, data
sources and timing, and the rules that govern data use
and data transformations.

7. The Directory Manager can:
 Import business models from CASE tools.
 Import metadata definitions from the prism
Warehouse Manager.
 Export metadata into catalogs, dictionaries, or
directories of many DSS access tools.
 Create flexible customized views based on end-
user requirements using graphical front-end
application.

8. User Sophistication levels
 Data warehousing is a relatively new phenomenon,
and a certain degree sophisticated is required on the
end users part of the warehouse.
 A typical organization maintains differ levels of
computer literacy and sophisticated within the user
community.
 The users can be classified on the basis of their skill
level in accessing the warehouse.

9. Casual users:
 These users are most comfortable retrieving information
from the warehouse in predefined formats , and running
preexisting queries and reports.
 These users do not need tools that allow fro sophisticated
and hoc query building and execution.

10. Power users:
 These users typically combine pre-defined queries
with some relatively simple and hoc queries that they
create themselves.
 These users can use drill-down queries to analyze the
result of the queries and reports.
 These users need access tools that combine the
simplicity of predefined queries and reports with a
certain degree of flexible.

11. Experts:
 These users tend to create their own complex queries
and perform a sophisticated analysis on the information
they
retrieve from the warehouse.
 These users know the data tools, and database well
enough to demand tools that allow for maximum
flexibility and adaptability.