This type of plan more has need than others of the personalizetions of the customers, beyond that of the ideas of the planner.
For hour the plan of the MV is still to the stage begins them, therefore the present job has been limited to imagine the needs informed you of who will have to manage supplyings, the conveyors, the supplyes of the warehouses and therefore via, being based on the detailed lists demanded from who it is carrying ahead this plan, that is the Unitec.
CHOSEN OF THE ARCHITECTURE AND THE TIPOLOGY
The choice that is intentional to carry ahead in this job of thesis has been that one to plan a
date mart, rather than a Datawarehouse in its thoroughness.
The choice has been moreover obliged from the same fact that
the Virtual Warehouse is taken care of a specific business function, that is the Logistics and Supplyings.
For Datamart
a under together agrees or an aggregation of gives to you, containing with of the important information for one particular
area of the business, one particular division of the company, one particular category of subjects.
The chosen architecture
for the model is an architecture to two levels.
In the truth of the plan in analysis the levels are four:
- Level of Sources
- Level of Feeding
- Datamart Level
- Level of Analysis
Graphically the passages are of easy understanding, like can be seen from the figure.
- The level or level of sources, in our case is only that one of the database operational, that is the DB of the MV. The possibilities to use other sources of give to you are immense; they could be the systems ERP, the systems legacy, give to you cartacei, Excel sheets.
- II the level or level of the feeding, allows to extract gives to you memoryzation to you in sources, often between heterogenous they, and to render them comprehensible to an only system, that is the Datamart. These instruments are defined like ETL (Extraction Transformation and Loading), and allow to integrate heterogenous outlines, let alone to extract, to transform, to clean up and to leak they give you of sources. The ETL can be implement you to the inside of the same company, or acquire to you to part on an immense market. In the case in examination it has not been implemented this phase.
- III the Level is that one of the Datamart. The information come collections in a single container. To flank to it it exists the container metadati that maintain information on the source, that is the structure of the original tables and the correspondences with those of arrival of the Datamart, information on the mechanisms of access, the procedures of clean and therefore via.
- IV the Level or level of analysis allows the efficient consultation and flexible of they give integrated to the ends of the drawing up of Report, analysis, simulation. From the technological point of view some technical abilities are demanded and logical to the customers who allow it to carry out an analysis through gives combines to you to you. The complex analyses gradually come yields little complex to comprise and to carry out, thanks to the propensione of the junior clerks you of this type towards the shape user-friendly, that is oriented to I use from part of the greater part of tipology customer, above all that one with insufficient computer science acquaintances.
The choice carried out in this job is that one to adopt a system ROLAP.
This idea very is motivated from the fact that has been carried out an enormous job in literature on the relational model (the same one of the Database of the MV) and that it is the used system more to business level; this large one I use in business within implies a greater acquaintance of I use and the administration.
However, the relational model does not include the concept of dimension, measure and hierarchy, that they are typical of architectures MOLAP, and that they are to the base of the multidimensional analyses.
In order to exceed this problem specific tipology of outlines are used that allow to traslare the multidimensional model on mattoni the base constituted from attributes, relations and ties of integrity.
This role is carried out from a particular type of outline, used in the present job, that is the outline to star or star outline.
The main problem of these systems is that one of the performances, that they suffer from the necessity to execute numerous operations of connection (join) on the tables that usually are of dimensions high.
The solution to this problem is that one of denormalizzare the outlines of departure in function of the volume of gives to you uses you and from the frequency of I use and therefore to rewrite the same ones they give more times to you in same database (the redundancies), with consequent increase of the used space, but also improvement of the access performances.
From a architetturale point of view, the ROLAP adoption demands of having an intermediate stage (to middleware ago) that from interpreter between the relational Server, where the Datamart is present, and the final customer who is the so-called one front-end.
This role allows translate interrogations OLAP formulated from the customer and tradurle in instructions SQL for the Datamart.
In the present job this role is carried out from the package "Analysis" contained "Service of series" in SQL Server 2000.
If instead a system MOLAP had been chosen, sure the insert would have been more difficult, considering the sparsity of commercially reperibili instruments, beyond that a greater difficulty of planning on which the literature is decidedly more miser.
A sure advantage would have been what the multidimensional operations are realizable in simple and natural way, without necessity to rerun to expensive complex and (in terms of performances) associations between tables, just because these systems are conceived in multidimensional way, ad hoc for the analysis. The performances are therefore optimal.