myITforum.com : Ascential Datastage and Business Objects software review

Hi friends !!! This month articles queue I will open from two software products description: Ascential Datastage and Business Objects.
My husband, Alexzander Nepomnjashiy in his “OLAP and Datawarehouse” series discuss typical datawarehouse conseptual model … As he wrotes: “…at the abstraction highest level, the architecture of solution of data warehouse is quite simple: the server (for example, Datastage - we shall give the description of it later) extracts the data from all necessary sources, groups and converts them according to pre-defined rules, and then loads in the warehouse. …” and “…For data access, comprising, storing and distribution of reports the client application is applied, permitting to make the on-line analysis and comprise reports, on the basis of data form the warehouse (for example, Business Objects – its description is also given later) …”

So let’s jump right into this software package reviewing . . .

Datastage software represents the set of products of Ascential Software corporation (former Informix Business Solutions), intended to provide support of transposition of the data from heterogeneous data sources into the storage, that is it is used to develop analytical systems for the level of the enteprise.

So, Ascential Datastage is the tool of visual development of procedures of extraction, conversion, loading (ETL - Extraction, Transformation, Loading) and control of the process of data transformation. Datastage peculiarities are the spanning of a wide spectrum of the data sources, maximum usage of possibilities of parallel calculations and reaching of extreme level of productivity without detriment to functionalities.

The description of diagrams of ETL-processes in Ascential Datastage is carried out with the help of Datastage Designer unit. Each of units of the diagram (also called stage) determines an operation on extraction of the data, their conversion or loading. The description of the ETL-procedure implies dragging buttons from a toolbar, appropriate to defined operation and the description of its properties. The line-up of units of the diagram determines one task (in terms of Ascential Datastage it is called job).

Within one job data reading from several sources and recording of the converted data in several receivers may be made simultaneously. Thus, both sources, and receivers may be located in various databases or files and have several copies in the network. Besides SQL for acceleration of the process of a data loading it is possible to use direct input / output in the file or so-called "fast loaders" (bulk loaders). Relational databases, flat files, databases on mainframes, Web-servers log files, XML-files may appear as the data sources. Receivers may have absolutely the same formats. Special resources of data extraction and loading are offered to simplify operation with inherited applications for most popular business applications, for example, such as SAP, Baan, PeopleSoft, Siebel.

Datastage nucleus uses multiprocessing environment for automatic allocation of the stream of independent jobs into several processes. This feature guarantees effective utilization of accessible operational resources and accelerates the speed of full data processing. Datastage server analyzes jobs before the beginning of their execution, selecting dependent operations and separately marking independent data flows. For each job Datastage nucleus generates some parallel processes, which are launched simultaneously (and may be fulfilled parallel on multiprocessing systems).

Prominent feature of Datastage system is maximum usage of the RAM and refusal of using (as temporary intermediate storage) a hard disk - everywhere where it is only possible. Due to this, intensity of operations of input / output, which is a bottleneck for productivity of any computing system, is considerably reduced.

Sorting and aggregation operations are carried out by Datastage system without addressing disk memory. Datastage carries them out during conversion, not using intermediate sources and SQL-operators. Datastage allows sorting in any point during conversion: before or after calculations, aggregation, on several ways of conversion (if recording is made in several receivers). Operations on conversion are carried out during passing lines on the way of conversions. During the phase of data conversion the user has a possibility to make graphics mapping ”a column in a column”, to create arithmetic and linear expressions for the description of on-off modulations with the data and to apply the firmware or user's functions.

The firmware resources of automatic creation and loading dynamic hash-tables in memory also provide improvement of productivity, significantly improving efficiency of search. Hash-tables may be loaded directly in execution time of the job (where search is required) or earlier and saved for usage hereinafter.

Datastage supports technologies of the high-speed data loading offered by Ascential partners and other suppliers of repositories. The loader of large scale arrays of the data or utilities “fast loading” ensures a high-speed insertion of line into the relational table, usually disabling support of logs of transactions (logging) and other service operations with transactions, which are used in registering applications.

Decision support systems require periodic data upgrade. Time required for upgrade, is defined by such parameters as: frequency of upgrade, availability of data receivers, possibility of access to the sources, expected size of the changed data. CDC technology (Change Data Capture) allows to change the data in storage directly at detection of a modification of the data in a source. CDC describes methods and the tools required for collection, modification and transition of recently entered or refreshed recordings. CDC support in Datastage reduces the loadtime required for repositories upgrade.

And in the conclusion

The set of Datastage products has a wide spectrum of the features necessary for transposition of the data into storage: their combining, filtering, coercion to the uniform format, elimination of backup, permission of multiple definitions for the same business - terms and many other things. Some Datastage products ensure visual designing of the process of transposition of the data and job controls on transposition, handling metadata and on their basis - handling of sources and data receivers. Distinctive, outstanding features and advantages of Datastage software are multy platform support, ability to operate huge data arrays, possibility of operation in the distributed computing environments, and also integration with systems of ERP class.

Business Objects

Business Objects company is the manufacturer of the integrated data access devices, the analysis of the information and construction of corporate analytical systems of the reporting. Business Objects unit, which supports access to various types of the data sources, including relational, OLAP-servers, and personal (“plane“) files, is the core one in the product line of the company. Business Objects allows uniting the information from the various data sources in one document or in one table, a graphics or for example, the diagram. For example, users may create a report, in which the comparative data on the real expenditures (from a corporate financial data base) and the information on planned expenses (located on the OLAP-server) will be displayed.

As a rule, analytical systems are meant for the broad audience of users among which there may be people far from IT problems. Therefore it is important, that possibilities for independent compilation of searches and representations of the information in convenient for the analysis sort were provided in such systems. Business Objects uses the concept of “semantic stratum” for this purpose, which includs dictionaries of business – terms. With their help complicated physical data model is mapped into understandable to the user objects of business. Each of the dictionaries can be customized on the defined data source and meant for the defined user group.

Business Objects user interface comprises diverse sets of pre-set commands for documents creation. Formulation of searches is reduced to choice and arranging of objects on a step-by-step board of searches. Formatting by a principle of towage simplifies creation of complicated complex reports.

All documents created in Business Objects may be stored in uniform repository, organized on the basis a relational DBMS, that simplifies a system control and troubleshoots safety by transmission and data storage. For data and system functions access processes of authorization and authentification of each user are supported. In Business Objects repository may be stored documents of other applications, for example such as Adobe Acrobat, Microsoft Excel, Microsoft Word.

Repository control is carried out with the help of Business Objects Supervisor unit, permitting to inspect users and operational resources in the distributed environment from one point. The object model of the safety system enables to install and change access rights both for groups, and for separate users. The administrator has the possibility to determine what function of the software, information and documents will be accessible to users /user groups. Supervisor also allows to use inherited and used by the Customer safety systems, including authentification of operating systems of the set of Microsoft Windows NT, LDAP or the usual system of passwords check for databases. As a result there is a possibility to set the uniform procedure of authentification for the heterogeneous data sources.

Broadcast Agent documents server is used for the job set up and conditions of automatic upgrade and distribution of documents. It has the distributed component architecture and supports various technologies of information transmission, including the e-mail, fax, pager, phone and PDA.

Main functions and the architecture of Business Objects software

The main functions achieved with the help Business Objects software are: 1) data access, 2) data analysis and 3) allocation of the information.

As we have already discussed above - Business Objects ensures data access of the end users not familiar with technical aspects of a used DB. A system nucleus is the semantic stratum in which the expert in the field of a DB describes rules of mapping of relational structure into object model. The end user, the analyst or the chief, using these data dictionaries, can independently build any searches to the various data sources, create reports, apply resources of visualization (OLAP) directly to sampling in the report. Business Objects is a corporate system, i.e. there are the resources, ensuring safety of usage of the information by various users, their collective operation and simple handle by the administrator.

Besides the version for a local network, there is as well the software version for web that allows to create corporate intranetwork and to connect subsidiaries of a company. Business Objects may be used for obtaining information directly practically from any data sources - starting from personal (plane) files up to relational databases with any architectural model, and also may function as a front-end resource for specialized multivariate databases.

Components of the system for the end user

Business Objects and WebIntelligence are the tools of user's creation of searches and constructions of reports. Accordingly these are "thick" and "thin" clients of the system. Business Objects and WebIntelligence unite in itself properties of powerful and at the same time intuitively understandable construction tool of complicated, professionally made out printed reports. In such reports may be united the data from different sources (including sources of different types), may be done complicated calculations and OLAP operations. WebIntelligence is a server-based application, installed on the server of Windows NT. It allows involving functionality Business Objects at zero costs for the end user.

Having on the computer only WEB - a browser, he/she can make arbitrary searches to relational databases of the corporation, carry out OLAP operations, and also view and update the data in the documents prepared in Business Objects. Both products use a common principle of data operation. The result of user's search is saved in the created document in the special multivariate format. The further operation with the document is made on the basis of sampling without addressing sources.

Database is connected only at modifications of searches or at upgrade of the report data. This kind of solution ensures rather flexible operation with relational DB and does not limit the user to frameworks of previously assembled data compared to traditional OLAP - servers. Along with analytical Business Objects platforms and WebIntelligence a Business Objects Set Analyser product for optimization of a velocity of searches to the database may be considered.

Business Objects - solution of a corporate scale

All users of Business Objects product within the enterprise can work in the uniform inspected and safe environment called the Repository. It is a set of the auxiliary tables, developed in any relational database on the server, accessible from all workstations. Repository is meant for the centralized storage of operational resources of Business Objects (semantic stratums and ready reports) and their exchange between users and developers.

Through documents repository jobs for automatic document processing by Business Objects Broadcast Server, which can carry out upgrade, printing, dispatch and other operations with documents, both under the schedule, and on response to events in the data sources are also transmitted. Repository is formed by administrative Business Objects Supervisor product. With the help of this product information system administrators assign rights of users on access to operational resources of Business Objects (Universe and to Ready documents), to the data of the database, functions of the system, and also set a valid operating schedule of the user.

Business Objects is the system open for programming. The majority of Business Objects products are OLE Automation servers. Therefore Business Objects may be used as the possible centralized, managed source of the analytical information for other, specialized applications.


		myITforum.com ©2010 \| Legal \| Privacy