0.1 Quizz 3 0.2 Quizz 4

may increase query performances -> VRAI. – In a Data Mart environment we use very often denormalized data model because we are querying very often on ...
Télécharger le PDF
39KB taille 5 téléchargements 344 vues
commentaire
Report
[SCIA] EPITA 2009

Data Warehouse

0.1

Quizz 3

– Performs Data loading into the databases where some transformations occur using sql and the database itself : ELT – Contains detailled and historical data. its model must be independant from the application (neutral model). It allows multi domain requests and bring the relevance to a customer relationship : Datawarehouse – Based on a data referential. allows bulk data load. all transformation occur prior to being loaded to database : ETL – Interfaces strongly coupled with the applications and the data warehouse. it must be able to manage almost real time events : ODS – Refers to data streams between one place to another place (file typem, server, location) and/or from one system to another : EAI

0.2

Quizz 4

– Model the Business implies a normalized data model -> Aujourd’hui on a besoin d’un schéma qui s’adapte aux évolutions des métiers et des entreprises -> 3 ème forme normale -> VRAI – Denormalized data model minimize the number of joins between tables and may increase query performances -> VRAI – In a Data Mart environment we use very often denormalized data model because we are querying very often on small or medium data tables -> FAUX (Oui on utilise des modèles dénormalisés, mais uniquement parcequ’on connait les requettes qui vont être faites) – Data Warehousing is a Coninually Evolving Process consisting in four main phases -> FAUX (Evolving process -> vrai, mais juste 3 etapes) – OLTP and Data Warehouse Workloads are very close ; consequently the data warehouse DBMS choice can rely on the OLTP DBMS choice -> FAUX (Module 3 pages 33/34), la vie entre la prod et le décisionnel est très différente. – In a Data Warehouse environment we may have a mixt of simple or complex pre-defined queries and simple or complex ad-hoc queries -> VRAI (Datawarehouse -> on répond a toutes les demandes de l’entreprise !) – The amount of detailed data volume as well as the number of concurrent users are the only parameters for choosing a Data Warehouse solution -> FAUX (Il faut aussi prendre en compte la complexité des requettes, ainsi que la complexité des modèles) – Scalability of the DW solution (Hardware and Software) is one of the central issues for DW project success -> VRAI -> 3 types d’architecture : 1

[SCIA] EPITA 2009

Data Warehouse

– PC domestiques – Clusters (pour backups, questions de sécurité) – Archi massivement parallèle – Data Volume size, Data Model complexity are the two elements to consider when selecting a Data Warehouse Solution -> FAUX (“the” est de trop)

0.3

Quizz 6

– Performs data loading into the databases where some transformations occur using sql and the database itself : ELT – Oltp process with multidimentional + relational databases, good solution for high scalability : HOLAP – Interface strongly coupled with the applications and the DataWarehouse. must be able to manage almost real time events : ODS – Dynamic analysis of multidimensional data. use a multidimensional denormalized data model determined by logical requirements : OLAP – Based on a data referential. allows a bulk data load. all transformation occur prior to being loaded to database : ETL – OLAP system based on relational tables : ROLAP – Traditional DataWarehouse plus ... very current detailed data integrated with historical data for strategic tactital and event driven business decision making. Allows timely updates - close to real time - as well as short, tactical queries that return in seconds : Activte DataWarehouse

2