- Information
- AI Chat
OLAP Server Architectures – Rolap versus Molap versus Holap
Principles Of Data Mining. (CSIT 440)
Montclair State University
Recommended for you
Students also viewed
Preview text
OLAP Server Architectures – ROLAP versus MOLAP versus HOLAP
Logically, OLAP servers deliver multidimensional data from data marts or warehouses to business customers without worrying about how or where the data are stored. However, OLAP servers' physical architecture and implementation must take data storage concerns into account. The following are some examples of OLAP processing implementations using a warehouse server: Relational OLAP (ROLAP) servers are the intermediary servers that sit between client front-end tools and a relational back-end server. They manage and store warehouse data using a relational or extended-relational DBMS, and they use OLAP middleware to support any gaps. ROLAP servers have aggregate navigation logic implementation, optimization for each DBMS back end, and extra tools and services. In general, ROLAP technology is more scalable than MOLAP technology. For instance, Microstrategy's DSS server uses the ROLAP methodology. Servers that offer multidimensional data views through array-based multidimensional storage engines are known as multidimensional OLAP (MOLAP) servers.
They directly correspond multidimensional perspectives to the array architecture of data cubes. Utilizing a data cube has the benefit of enabling quick indexing of precomputed summary data. Be aware that if the data set is sparse, storage utilization for multidimensional data stores may be low. Techniques for sparse matrix compression should be investigated in such circumstances. To manage dense and sparse data sets, many MOLAP servers use a two-level storage representation: Sparse subcubes use compression technologies for effective storage consumption, whereas denser subcubes are detected and stored as array structures. Servers for hybrid OLAP (HOLAP): The hybrid OLAP approach combines MOLAP and ROLAP technology, taking advantage of MOLAP's quicker processing speed and ROLAP's larger scalability. For instance, a HOLAP server might enable the storage of substantial amounts of specific data in relational databases while keeping aggregations in a separate MOLAP store.
A hybrid OLAP server is supported by Microsoft SQL Server 2000. SQL servers that specialize: Some database system suppliers develop dedicated SQL servers that offer advanced query language and query processing capabilities for SQL queries over star and snowflake schemas in a read-only environment to fulfill the growing demand for OLAP processing in relational databases. What kind of data storage does ROLAP and MOLAP designs use in practice? First, let's examine ROLAP. ROLAP, as the name suggests, stores data in relational tables for online analytical processing. Remember that a base fact table is the fact table associated with a base cuboid. Data is stored in the basic fact table at the level of abstraction denoted by the join keys in the specified data cube's schema. Fact tables, sometimes known as summary fact tables, can be used to hold aggregated data. A few summary fact tables keep both the data from the basic fact tables and the aggregate data (see Example 3).
To keep just aggregated data, distinct summary fact tables can be utilized for each abstraction level. The data cube can be thought of conceptually as a form of multidimensional data generalization. In general, data generalization reduces the number of dimensions to summarize data in concept space with fewer dimensions, or it replaces relatively low-level values (such as numerical values for an attribute age) with higher-level concepts (such as young, middle-aged, and senior) (e., removing birth date and telephone number when summarizing the behavior of a
group of students). It is helpful to be able to explain topics in precise and succinct language at generalized (rather than low) levels of abstraction since databases store a lot of data. Users can examine the general behavior of the data more easily when data sets may be generalized at several levels of abstraction. Given the AllElectronics database, for instance, sales managers may prefer to view the data generalized to higher levels, such as summarized by customer groups according to geographic regions, frequency of purchases per group, and customer income, rather than looking at individual customer transactions.
As a result, we get at the idea of concept description, a type of data generalization. Typically, a notion refers to a data set, such as a list of frequent consumers, graduate students, etc. Concept description is not just an enumeration of the data when it comes to data mining tasks. Concept description, on the other hand, produces descriptions for characterisation and comparison of facts. When the notion to be expressed pertains to a class of objects, it is also referred to as a class description. While concept or class comparison (also known as discrimination) gives descriptions comparing two or more data collections, characterization provides a brief and succinct summary of the specific data collection. Up till now, we have researched multidimensional, multilevel data generalization techniques in data warehouses employing data cube (or OLAP) approaches to idea definition. How well-suited is data cube technology for massive data sets' various concept definition tasks? Think about the following scenarios. Types of complex data and aggregation: The foundation of data warehouses and OLAP technologies is a multidimensional data model, which presents data as a data cube made up of dimensions (or attributes) and measurements (aggregate functions). But many modern OLAP systems limit measures to numeric data and dimensions to non-numeric data.
In practice, the database may contain attributes of several data types, such as text, images, spatial, numeric, and non-numeric attributes, all of which ought to be mentioned in the concept description. Additionally, the collection of non-numeric data, the combining of spatial regions, the composition of images, the integration of texts, and the grouping of object pointers are examples of advanced data kinds that may be included in the aggregate of attributes in a database. OLAP, with its limitations on the types of dimension and measure that are feasible, thereby constitutes a streamlined paradigm for data analysis. When necessary, concept descriptions should be able to manage the complicated data types of the attributes and their aggregations.
Automation versus user control Data warehouses' online analytical processing is a user- controlled procedure. Users are generally in charge of choosing the dimensions and applying OLAP operations (such as drill-down, roll-up, slicing, and dicing). Although most OLAP systems have controls that are quite user-friendly, users still need to have a solid understanding of what each dimension does. Users may also need to describe a lengthy list of OLAP operations in order to find a good description of the data. In order to provide an engaging summary of the data, it is frequently desirable to have a more automated approach that aids users in deciding which dimensions (or qualities) should be included in the analysis and the extent to which the provided data set should be generalized. This section describes attributeoriented induction, a
OLAP Server Architectures – Rolap versus Molap versus Holap
Course: Principles Of Data Mining. (CSIT 440)
University: Montclair State University
