Jens Albrecht

Normal forms for multidimensional databases

In the area of online analytical processing (OLAP), the concept of multidimensional databases is receiving much popularity. Thus, a couple of different multidimensional data models were proposed from the research as well as from the commercial product side, each emphasizing different perspectives. However, very little work has been done investigating guidelines for good schema design within such a multidimensional data model. Based on a logical reconstruction of multidimensional schema design, this paper proposes two multidimensional normal forms. These normal forms define modeling constraints for summary attributes describing the cells within a multidimensional data cube and constraints to model complex dimensional structures appropriately. Multidimensional schemas compliant to these normal forms do not only ensure the validity of analytical computations on the multidimensional database, but also favor an efficient physical database design.

Genetic relationships between bovine herpesvirus 4 and the gammaherpesviruses epstein-Barr virus and herpesvirus saimiri

The overall arrangement of genes in the unique central part of the bovine herpesvirus type 4 (BHV-4) genome has been deduced by analysis of short DNA sequences. Twenty-three genes conserved in at least one of the completely sequenced herpesviruses have been identified and localized. All of these genes encoded amino acid sequences with higher similarity to proteins of the gammaherpesviruses Epstein-Barr virus (EBV) and herpesvirus saimiri (HVS) than to the homologous products of the alphaherpesviruses varicella-zoster virus and herpes simplex virus type 1 or the betaherpesvirus human cytomegalovirus. The genome organization of BHV-4 had also an overall colinearity with that of the gammaherpesviruses EBV and HVS. Furthermore, the BHV-4 genes content and arrangement were more similar to those of HVS than to those of EBV, suggesting that BHV-4 and HVS are evolutionarily more closely related to each other than either are to EBV. BHV-4 DNA sequences were generally deficient in CpG dinucleotide. This CpG deficiency is characteristic of gammaherpesvirus genomes and suggests that the BHV-4 latent genome is extensively methylated. Despite several biological features similar to those of betaherpesviruses, BHV-4 displays the molecular characteristics of the representative members of the gammaherpesvirinae subfamily.

Management of multidimensional aggregates for efficient online analytical processing

Proper management of multidimensional aggregates is a fundamental prerequisite for efficient OLAP. The experimental OLAP server CUBESTAR whose concepts are described, was designed exactly for that purpose. All logical query processing is based solely on a specific algebra for multidimensional data. However, a relational database system is used for the physical storage of the data. Therefore, in popular terms, CUBESTAR can be classified as a ROLAP system. In comparison to commercially available systems, CUBESTAR is superior in two aspects. First, the implemented multidimensional data model allows more adequate modeling of hierarchical dimensions, because properties which apply only to certain dimensional elements can be modeled context-sensitively. This fact is reflected by an extended star schema on the relational side. Second, CUBESTAR supports multidimensional query optimization by caching multidimensional aggregates. Since summary tables are not created in advance but as needed, hot spots can be adequately represented. The dynamic and partition-oriented caching method allows cost reductions of up to 60% with space requirements of less than 10% of the size of the fact table.

Query optimization by using derivability in a data warehouse environment

Materialized summary tables and cached query results are frequently used for the optimization of aggregate queries in a data warehouse. Query rewriting techniques are incorporated into database systems to use those materialized views and thus avoid the access of the possibly huge raw data. A rewriting is only possible if the query is derivable from these views. Several approaches can be found in the literature to check the derivability and find query rewritings. The specific application scenario of a data warehouse with its multidimensional perspective allows the consideration of much more semantic information, e.g. structural dependencies within the dimension hierarchies and different characteristics of measures. The motivation of this article is to use this information to present conditions for derivability in a large number of relevant cases which go beyond previous approaches.

Set-Derivability of Multidimensional Aggregates

A common optimization technique in data warehouse environments is the use of materialized aggregates. Aggregate processing becomes complex, if partitions of aggregates or queries are materialized and reused later. Most problematic are the implication problems regarding the restriction predicates. We show that in the presence of hierarchies in a multidimensional environment an efficient algorithm can be given to construct - or to derive - an aggregate from one or more overlapping materialized aggregate partitions (set-derivability).

Aggregate-based query processing in a parallel data warehouse server

In the last years data warehousing has emerged as a fundamental database technology providing the basis for online analytical processing (OLAP). In general, analytical queries involve aggregations of large data sets. This results in serious performance problems if ad-hoc queries are to be answered online. One method to avoid performance bottlenecks is to use parallel hardware, i.e. SMP or MPP machines which are able to cope with the data volume. Another optimization approach specific to data warehousing is to preaggregate some of the results in order to avoid scanning the base relations. The prototypical OLAP system CUBESTAR PARALLEL SERVER combines both approaches. In order to achieve high query performance with low hardware costs, we present a technique for the dynamic, i.e. query-behavior and load-dependent, use and management of multidimensional aggregates in a shared-nothing workstation cluster.

Building a real data warehouse for market research

This paper reflects the results of the evaluation phase of building a data production system for the retail research division of the GfK, Europes largest market research company. The application specific requirements like end-user needs or data volume are very different from data warehouses discussed in the literature, making it a real data warehouse. In a case study, these requirements are compared with state-of-the-art solutions offered by leading software vendors. Each of the common architectures (MOLAP, ROLAP, HOLAP) was represented by a product. The result of this comparison is that all systems have to be massively tailored to GfKs needs, especially to cope with meta data management or the maintenance of aggregations.

Data Mining in a Multidimensional Environment

Data Mining and Data Warehousing are two hot topics in the database research area. Until recently, conventional data mining algorithms were primarily developed for a relational environment. But a data warehouse database is based on a multidimensional model. In our paper we apply this basis for a seamless integration of data mining in the multidimensional model for the example of discovering association rules. Furthermore, we propose this method as a userguided technique because of the clear structure both of model and data. We present both the theoretical basis and efficient algorithms for data mining in the multidimensional data model. Our approach uses directly the requirements of dimensions, classifications and sparsity of the cube. Additionally we give heuristics for optimizing the search for rules.

Using Semantics for Query Derivability in Data Warehouse Applications

Materialized summary tables and cached query results are frequently used for the optimization of aggregate queries in a data warehouse. Query rewriting techniques are incorporated into database systems to use those materialized views and thus avoid accessing the possibly huge raw data. A rewriting is only possible if the query is derivable from these views. Several approaches can be found in the literature to check the derivability and find query rewritings. However, most algorithms either find rewritings only in very restricted cases or in complex cases which rarely occur in data warehouse environments. The specific application scenario of a data warehouse with its multidimensional perspective allows the consideration of much more semantic information, e.g. structural dependencies within the dimension hierarchies and different characteristics of measures. The motivation of this article is to use this information to present simple conditions for derivability in a large number of relevant cases which go beyond previous approaches.

Optimistic Coarse-Grained Cache Semantics for Data Marts

Data marts and caching are two closely related concepts in the domain of multi-dimensional data. Both store pre-computed data to provide fast response times for complex OLAP queries, and for both it must be guaranteed that every query can be completely processed. However, they differ extremely in their update behaviour which we utilise to build a specific data mart extended by cache semantics. In this paper, we introduce a novel cache exploitation concept for data marts - coarse-grained caching - in which the containedness check for a multi-dimensional query is done through the comparison of the expected and the actual cardinalities. Therefore, we subdivide the multi-dimensional data into coarse partitions, the so called cubletets, which allow to specify the completeness criteria for incoming queries. We show that during query processing, the completeness check is done with no additional costs