luckyhillockData Management

Nov 29, 2012 (5 years and 6 months ago)


Survival guide for information system

Databases (Bases de données)

: I think that everyone know more or less what is a database but I will
precise what are the database

that are used in the information system
. Obviously, no one use
Access in an informati
on system (well in fact no one should use it in an information

…), the
databases that are used for these system

are bigger and more expensive.

There are four main actors on the big database market :

Oracle from the Oracle company

DB2 from IBM

Sql S
erver from Microsoft

Sybase Adaptive Server Enterprise

from Sybase

You may also find some free alternatives :

PostgreSql that compete with the four big actors

Mysql that is smaller but that is used a lot for small/medium usages

E.R.P. (Enterprise Ressource


E.R.P. en français)
: An E.R.P. is the core of the information
system of an enterprise. It regroups all the data gathered from every department

of the firm. The
data are entered
and consulting
using generally an

interface that look lik
es an access form.

The E.R.P. also realize specific computations (generally when everyone is sleeping) in order to
compute the input data (for example to consolidate the data).

The E.R.P. are generally very expensive for big enterprises but now, as every b
ig enterprises have an
E.R.P. , the editors now focus on the SME with cheaper E.R.P. . They are also free E.R.P. that could be
used on the S.M.E. market.

Generally the installation of an E.R.P. is very long and necessitate more than one year to be used by
the business staff. For example, the installation of SAP to Renault took 2 years and used 200 people.
The E.R.P. are modular and the enterprise choose which module to buy depending of it’s need, then
the module are modified to adapt to the need and specifi
c module

are developed inside the
company to completely fit to the need.

They are 4 main actors on the market of the E.R.P. :

S.A.P. with SAP. ECC / S.A.P. R/3 for the biggest enterprises and SAP business one for the

Oracle with Oracle ebusines suit

(but as they acquired a lot of

companies they
propose a lot of different E.R.P. such as PeopleSoft or
JD Edwards EnterpriseOne
). In the
future, Oracle will only sell Oracle fusion that is a merge of all the E.R.P. they own.

Microsoft with
Microsoft Dynamics that is mainly conceived for the SMEs even if they try to
sell it to the biggest enterprises.

Sage with Sage ERP X3 (and some other ones for smaller companies) that is also dedicated
for SMEs

PeopleSoft (one of the best looking E.R.P.

Business Intelligence
(Informatique décisionnel):
The business intelligence regroups all the tools
that can be used to analyze the huge amount of data that are collected by the information system
(and especially by the E.R.P.). They includes the repor
ting tools, dashboards, data cube etc… I will
explain these terms just after.

The biggest B.I. companies are :







A few years ago they were numerous enterprises but they were all bought by one the main actors.

warehouse (Entrepôt de données)

A data warehouse is
a database where the data are
centralized and computed in order to realize

on it.

Generally the data warehouse are very
big (the biggest have a capacity of more than 20


000 GB(GigaOctets)
) because the goal is to
compute everything before the user could access to the data using business intelligence tool such as
data cube.

Data mart (Magazin de données) :

A datamart is a subset of a data warehouse
that is specific to one
subject in order to

analyze one precise point.

Data Cube
(Cube de données)

A data cube is
a tool that can aggregate the data depending on fields.
For example a data cube could list the sales per country, category and business unit during the
month of November 2007. It is po
ssible to create a data cube in excel using the pivot table tool
(tableau croisé dynamique).

Example of a data cube that

warranty reserve

Dashboards (tableaux de bord)
: The dashboards are tools that can be used by the top management
in ord
er to see the evolutions of the K.P.I (Key performance indicators) that are benchmark in order
to measure the performance of the enterprise. The dashboards are generally made with a set of

Reporting (Reporting)
The reporting tools (like some part
of Cognos or Sql Server

are used in order to build reports (which can be excel files, word files, web file etc…) that are the
results of the business intelligence analysis.

Data mining (Data mining)
The data mining regroups all the to
ols that are used in order

to retrieve
information from the huge volume of data gathered in the data warehouse. The data mining is
mainly used on three types of analysis :

Divide the data into groups (Example for the computer market : identify the differen
t types
of clients (professionals, geeks, family users etc…))

Recognize if an element is part of one group or another (To recognize if for example a
customer of mart that sale computer is a geek, a professional etc…)

Identify relationship between elements

(for example in the supermarket, to identify if a
customer which buy a television could also buy at the same time chips)

The different analysis used generally probability (for example the bayasian network), computer
science algorithm (neuronal networks ie

: artificial intelligence) or the set theory (clustering

The data mining is used a lot in marketing analysis.


Extract Transform and Load or E.L.T Extract Load and Transform :

The E.T.L. are tools that
are used to transfer huge volumes
of data from one source of data (generally a database but it could
be also an excel file, a text file etc…) to another. It is used a lot in enterprise in order to consolidate
the data and to feed the data warehouse.

The E.T.L. are mostly provided by the ed
itor of database (Microsoft, IBM, Oracle, Sybase etc…) but
other editors provide E.T.L. such as Informatica and Talend (that provide a free E.T.L.). If you want
more information on the E.T.L. (but I doubt that it may interest you), I could provide you
a mo
detailed document because I have made the scientific part of my engineer internship memory on this