INFORMATION TECHNOLOGIES OLAP
O.P.Il’yna[1] And R.R. Sharipov[2] St. Petersburg State University Of Economy And Finance, Russia

Structure of information systems for corporations includes Systems for Support of the Decisions (SSD), On-Line Analytical Processing (OLAP). The large enterprises put huge financial assets in creation and support of the analytical systems allowing to carry out profound analysis of economic activity. Information systems of small and average business include processes of the registration information (system On-Line Transaction Processing - OLTP), the set of analytical functions is rather limited.

For tasks of strategic planning of activity of the enterprise it is necessary to supply:

1. Continuous account and analysis of the current results of activity.

2. Creation and conducting storehouses of the data for the long period.

3. Analytical data processing with use of methods of the statistical analysis and modeling;

4. Forecasting meanings of parameters of activity of the enterprise for prospect.

Most frequently in SSD the following kinds of analytical processing are applied:

the filtration of the given large files for their visual control.

the accounts and calculations of economic parameters to various attributes of the data.

the aggregating of economic parameters for calculation of statistical results (sum, average, maximum, minimum etc.).

the consecutive «a deepening» in the data (drill down) for their detailed elaboration (disaggregating of economic parameters).

the construction and analysis of trend’s lines of meanings of economic parameters etc.

As a rule, the construction of the reports requires the large computing capacities.

The basic source of the data for SSD becomes Data Warehouse - DW, containing the data, the storage area for meta-data; the interface means (Query Languages, Generators of the Reports and Methods of the Analysis etc.). At updating sources is made update of DW.

The developers of the software have begun realization of functions for analytical processing of the data. For example, program «1С: Предприятие» for automation of administrative activity of the small and average enterprises provides creation of the advanced reporting. By the basic problem for the analytical reports (for example, «Analysis of sales», «Analysis of Parties of the goods») is the large time of data processing (reaches hour and more at firms with a plenty of the goods and spent operations). The period of the analysis very strongly influences time of formation of the analytical reporting. For example, the report on movement of the goods for last day is done some seconds, for last year - more than 10 minutes. Even in SQL - versions the reports with use large выборок of the data are formed not in real time.

The second problem, which is urgent for the users «1С: Предприятие» - limited set of the analytical reports. Really, to receive the necessary data on purchases, sales, payments and debts is quite real, but the forms of the reports delivered in typical configurations, are inconvenient for the analysis that occurs in firm (for example, dynamics of the warehouse rests such as "Rest/day"). The built-in tools of development allow to create the new reports, but they are labour-consuming and difficult for the ordinary users. Such situation is a brake at acceptance of the administrative decisions. The software should be accessible to creation the analytical reports to the administrative personnel, demanding the minimum of time (only several minutes).

The retrospective analysis is carried out on the basis of several DB, kept on the periods of the account (year, quarter or month). These DB can differ by data structure, technique of the account etc. For example, «convolutions» of the rests are done, the documents before the certain date leave, the items of information on the rests of the goods, parties, accounts for the certain period are kept only.

The time of the report’s creation depends from DBMS: in the greater degree for the DBF-versions, to a lesser degree for the SQL-versions. During creation of the complex report input of the operative documents is slowed down. It is desirable to share processes of reception of the reports from transaction registration. During the reports creation it is impossible to carry out structural data’s transformation for DB.

The OLAP-technologies provide effective access to DB of a various kind and format, high quality processing and price appeal to the small and average enterprises using mainly the program «1С:Предприятие» on base MS-SQL Server and the program MS Office.

The base tool of the analysis is the pivot tables, which allow:

to form the various reports without participation of the programmers;

to detail of the report up to the primary documents during the analysis;

to see the information in cuts of clauses and centers of the financial account etc.

 The OLAP-reports are formed in tens time of the faster usual reports that the total is formed beforehand. Even the very complex OLAP-reports usually are under construction no more than 5 seconds, while their analogues («1С: Предприятие») — 10–20 minutes.

In the company some DB and different configurations of system of the various suppliers are used at once (even if the same program, for example, «1С: Предприятие»). The OLAP-technologies allow consolidating DB, to differentiate access to the data at a level of the report or separate measurements of parameters. The information technologies OLAP are effective on a ratio/quality, for the average companies they cost $200- $ 500. OLAP-reports are easily modified by the users, one OLAP-report is equivalent to the whole set of the simple reports.

Unfortunately, use OLAP in registration system of a OLTP-type is complicated by the several factors:

1. The OLAP-quires demand a lot of time and essentially reduce productivity of OLTP-system.

2. The registration systems are non-uniform on structure and formats of the stored data given in them frequently incomplete and unmatched.

3. The retrospective analysis and construction of the trend’s lines sometimes is impossible, the information for the last periods may be lost at updating.

Besides there are also technological problems of introduction OLAP in registration systems: various hardware and program platforms (operational systems, DBMS, program languages, protocols of the binding software etc.), dispersed data storage etc.

The process of integration of OLAP-technologies and information systems can be carried out by a various image:

the creation of specific analytical means within the framework of OLTP-systems (most expensive way);

the creation of DW of the data on a basis DB OLTP-systems as large volume hypercube and use standard OLAP-technologies (less expensive way).