WORKSHOP ON MANAGING EVOLUTION OF DATA WAREHOUSES - MEDWa

co-located with the ADBIS conference
September 7, 2009
Riga, Latvia
Workshop program
KEYNOTE TALK 1

Alegros way from XLS based controlling to a modern BI environment
Christian Maar, CIO Allegro Group

Christian Maar (42) received his degree in Economics in Nuremberg (Germany). He used to work for the Siemens AG and led several complex software development projects. Afterwards Christian was the head of development and architecture in a SMB consulting firm called FMDdata. Further on he joined Oracle Cooperation and worked there as a principal consultant in the BI & Data Warehouse area. During that time he led several international projects. After leaving Oracle he worked for Ariba Cooperation worldwide and was in charge of the client relationship division in Central Eastern Europe Region. Then Christian joined IBM to work worldwide as a principal in the e-Procurement / Supply Chain area. After that he headed the eCommerce devision of KarstadtQuelles mail order business. Since 2007 Christian has been leading the Allegro Group in the position of CIO.

Abstract
Allegro is the most successful eCommerce company in East Europe and the biggest non-eBay auction platform worldwide. This year the Allegro group is celebrating their 10-year anniversary. The company is in the middle of a transition from a small mid size company to corporation with more than 1.000 employees. In all relevant areas things are becoming more and more complex. Traffic and business growth are not in sync anymore. Two times more internet traffic is needed to grow 100% on the business side. As a consequence of Allegros rapid growth over the years the amount of data/information is continually increasing. This presentation will give some inside out view of two important focus areas of the company: (1) Allegros project to built one of the most efficient and scalable state of the art DWH infrastructures in Poland or even in Central East Europe and (2) Allegros fraud detection and prevention mechanisms in the online auction area.
KEYNOTE TALK 2: will be given at ADBIS

Multidimensional Model: From Conceptual to Implementation Perspectives
Prof. Dr. Elzbieta Malinowski G., University of Costa Rica

Data warehouse and OLAP systems are widely required during the decision-support process, since they provide integrated data and organize it in a way that facilitates the expression of complex queries. However, the necessity of building DW systems that fulfill user expectations was ahead of formal approaches such as the ones we had for operational databases. This leads to the twofold situation: on the one hand, practitioners and commercial tools propose particular solutions for managing different multidimensional elements; on the other hand, scientific community focuses on different aspects, such as different types of hierarchies, dimensions, support for temporal changes, among other, without a clear harmonization to converge into a feasible solution acceptable by vendors and practitioners.
In this talk we refer to different kinds of hierarchies and dimensions already classified in our previous work; these elements exist in real-world applications and are required during the decision-making process. Many of these hierarchies and dimensions can already be implemented in commercial tools, e.g., in Microsoft SQL Server Analysis Services (SSAS). We will demonstrate the importance of using a conceptual multidimensional model to facilitate the process of understanding user requirements by characterizing different kinds of hierarchies and dimensions. We have chosen SSAS as an example of a commercial OLAP tool because, to our knowledge, it provides different kinds of hierarchies and dimensions that can be included in the cube without incurring into any programming effort, i.e., using a wizard or the click-and-drag mechanism. We will compare the representation of different kinds of hierarchies and dimensions in the conceptual model, in the relational logical model (star or snowflake schema), and in the SSAS OLAP cube designer.
In addition, we will also refer to the temporal extension of the conceptual multidimensional model and to the design methodology for temporal data warehouses. We believe that even though there is a significant research effort in managing temporal changes to dimension instances or/and to dimension structures under of the term temporal data warehouses, the separate research effort mainly focuses on implementation details. Therefore, on the one hand, users cannot express which temporal changes are important for their applications; on the other hand, it is difficult to agree upon a common framework for expressing and managing changes to data.
SCHEDULE

Date: September 7, 2009
Venue: Riga Technical University, Meza str. 1/3, 1/4


08:00-09:00
Registration

09:00-09:30 Opening (all workshops)

09:45-11:15
Keynote talk 1 (joint for MEDWa and Trust4Web)

Alegros way from XLS based controlling to a modern BI environment (Christian Maar, CIO Allegro Group)

11:15-11:45 Coffee break

11:45-13:15 Session 1: ETL Evolution and Evolving DW Architectures
  • Rule-based Management of Schema Changes at ETL sources (G. Papastefanatos, P. Vassiliadis, A. Simitsis, T. Sellis, Y. Vassiliou)
  • ETL evolution from data sources to datawarehouse using mediator data storage (A. Dolnik)
  • Multiversion Spatio-Temporal Data Warehouse Case Study (M. Gorawski, M. Gorawski)

13:15-14:30 Lunch

14:30-16:00 Session 2: Managing Versions of Data Warehouses 1
  • Detection and Propagation of Complex Evolutions in Multiversion Data Warehouse (I. Zouari, H. Boukhriss, F. Ghozzi and R. Bouaziz)
  • An Approach to Supporting Data Warehouse Schema Versions: Overview and Case Study (D. Solodovnikova)
16:00-16:30 Coffee break

16:30-17:30 Session 3: Managing Versions of Data Warehouses 2
  • Indexing Data Versions in Multiversion Data Warehouse by Means of Multiversion Bitmap Join Index (J. Chmiel)
  • Evolving Constraints in XML Data Transformation (MD. S. Shahriar, J. Liu)

17:30-18:30 Closing Discussion: Open issues and current trends in managing data warehouse evolution

19:30 A joint dinner for volunteers (unfortunately not sponsored :-((( )
WORKSHOP PROCEEDINGS
An extended post-worksop version of accepted and presented papers will be published in a Springer Verlag LNCS volume. Accepted papers will also be published in a local conference proceedings (available at the workshop).

WORKSHOP CHAIR
Robert Wrembel
Poznan University of Technology
Institute of Computing Science
Poland