Pages:     | 1 |   ...   | 40 | 41 || 43 | 44 |   ...   | 54 |

Because of this the narrow versions of such software are offered at the market. Unfortunately those versions are not as convenient as they are advertised and provoke many additional problems during the implementation process and exploitation.

Our approach is to build information complexes for information service of business accounting and decision making based on numbered information spaces [Markov, 2004a], which may support RTAP on the level of ODS Class I and this way to reduce the expenses for maintenance separate DW. This goal may be achieved using the FOI Archive Manager (ArM) .

FOI Archive Manager (ArM) The FOI Archive Manager (ArM) is a tool for building numbered information spaces. ArM is based on the MultiDomain Information Model (MDIM). It has been established more than twenty years ago. For a long period it has been used as a basis for organization of the information bases. The first publication which contains some details from MDIM is [Markov, 1984] but as a whole the model was presented in [Markov, 2004a]. There exist several realizations of FOI Archive Manager (ArM) for different hardware and/or software platforms. The newest ArM Version 9 for IBM PC developed using DELPHI for MS Windows XP is called ArM32.

Let remember the main possibilities of ArM32 [Markov, 2004b] using some definitions of MDIM.

Basic information element of DI is an arbitrary long string of machine codes (bytes). When it is necessary the string may be parceled out by lines. The length of the lines may be variable. In ArM32 the length of the string may vary from 0 (zero) up to 230 (1G) bytes. There is no limit for the number of strings in an archive but theirs total length plus internal indexes could not exceed 4G bytes in a single file.

Let E is a set of basic information elements: E = {e | e E, i=1,, m }.

1 1 i i 1 Let is function which defines biunique correspondence between elements of the set E and elements of the 1 set C of positive integer numbers: C = {c | c N, i:=1,, m }, i.e. : E C The elements of C are said to 1 1 i i 1 1 1 1. be number codes of the elements of E. The triple S = ( E,, C ) is said to be numbered information 1 1 1 1 space of range 1.

Information Systems The triple S = (E,, C ) is said to be numbered information space of range 2 iff E is a set which 2 2 2 2 elements are numbered information spaces of range 1 and is function which defines biunique correspondence between elements of E and elements of the set C of positive integer numbers:

2 C = {c | c N, j:=1,,m }, i.e. : E C 2 j j 2 2 2 2.

The triple S = (E,, C ) is said to be numbered information space of range n iff E is a set which n n n n n elements are information spaces of range n-1 and is function which defines biunique correspondence n between elements of E and elements of the set C of positive integer numbers: C = {c | c N, k:=1,,m }, n n n k j n i.e. : E C n n n.

The sequence A = (c,c,,c ) where cC, i=1,,n is called multidimensional space address of range n of n n-1 1 i i a basic information element. Every space address of range m, m

Every index may be considered as basic information element, i.e. as a string, and may be stored in a point of any information space. In such case it will have a multidimensional space address which may be pointed in the other indexes and, this way, we may build a hierarchy of indexes. So, every index which points only to indexes is called metaindex.

Let G = {S | i=1,...,m} is a set of numbered information spaces.

i Let ={ : S S | i=const, j=1,m} is a set of mappings of one main numbered information space S G, ij i j i i=const, into the others S G, j=1,m, and, in particular, into itself. The couple: = (G, ) is said to be an j aggregate.

The ArM32 elements are organized in numbered information spaces with variable ranges. There is no limit for the ranges the spaces. Every element may be accessed by correspond multidimensional space address (coordinates) given via coordinate array of type cardinal. At the first place of this array the space range needs to be given. So, we have two main constructs of the physical organizations of ArM32 numbered information spaces and elements.

The main ArM32 operations with basic information elements are: ArmRead (reading a part or a whole element);

ArmWrite (writing a part or a whole element); ArmAppend (appending a string to an element); ArmInsert (inserting a string into an element); ArmCut (removing a part of an element); ArmReplace (replacing a part of an element); ArmDelete (deleting an element); ArmLength (returns the length of the element in bytes).

The ArM32 numbered information spaces are ordered and main operations within spaces take in account this order. So, from given space point (element or subspace) we may search the previous or next empty or non empty point (element or subspace). In is convenient to have operation for deleting the space as well as for count its nonempty elements or subspaces.

The ArM32 logical operations defined in the multi-domain information model are based on the classical logical operations - intersection, union and supplement, but these operations are not so trivial. Because of complexity of the structure of the spaces these operations have at least two principally different realizations based on codes of information spaces elements and on contents of those elements.

The ArM32 information operations can be grouped into four sets corresponding to the main information structures: elements, spaces, aggregates, and indexes. Information operations are context depended and need special realizations for concrete purposes. Such well known operations are, for instance, transferring from one structure to another, information search, sorting, making reports, etc.

At the end there exist several operations which serve information exchange between ArM32 archives (files) such as copying and moving spaces from one to another archive.

Fourth International Conference I.TECH 2006 ArM32 engine supports multithreaded concurrent access to the information base in real time.

Very important feature of ArM32 is possibility not to occupy disk space for empty structures (elements or spaces).

Really, only non empty structures need to be saved on external memory.

Complex FOI Complex FOI is an integrated software environment for economical information processing and business analysis. The main features of Complex FOI [Markov et al, 1994] are built on three levels, which correspond to the Pyramidal Information Model (PIM) presented in [Markov et al, 1993]. The levels of this model are Strategy, Analysis, and Service. Every level contains three parts, which correspond to Human Resources, Materials, and Finances of the enterprise. It easy to see that there exist correspondence between PIM and ODS and DW.

The main set of concrete systems for information processing is included on "Service" level. They are aimed to service the operative work and control. For instance, there exist systems for service the enterprise financial tasks such as computing of salaries [Markov et al, 1996a], systems for managing different material stores using appropriate information access - by names or by numbers of goods [Markov et al, 1995a], systems for maintenance of fixed assets [Markov et al, 1996b], etc. An example of another class of service systems is one for automated payment of consumption of water and other communal services in a town as well as the specialized service systems, such as one for computing the price of building of some architectural object. It is clear, the legacy applications of the enterprise are assumed to be on this level too.

All these systems are integrated with the upper level (Analysis) via very convenient interface the natural language standard accounting records which are the usual transaction form for accounting process. Furthermore, the information in Complex FOI is distributed in correspond numbered information spaces in accordance to usual every day financial accounting information structures. This make integration possible and automated information exchange is simple and comprehensible.

There is only one system on level "Analysis". It is an ODS with possibilities for accounting as well as for account analysis [Markov et al, 1995b]. This is the main tool for enterprise financial control and managing which support automated day-to-day operations (purchasing, banking etc), transactions access and modifying a few records at a time, application oriented database design, and metric: transactions/sec. The main structure of this level is the financial ledger - usually it is a numbered information space of range up to 10. Its subspaces represent accounting divisions, groups and accounts, as well as sub-accounts on several sub-levels. Every space may contain operational and historical data in the same time.

The main feature of the level "Strategy" is the decision support. All information from low levels can be used for supporting the processes of business decisions in the group of leaders of the enterprise. The functionality of this level covers the usual understanding of data warehouse but it is realized as distributed RTAP engine which support complex queries that access records with operational and/or historical data for trend analysis.

Because of special multidimensional organization, in Complex FOI the analytical pre-computation can be provided in real time during the operative work and its results (elements, spaces, aggregates, and indexes) can be stored in corresponded structures of the multidimensional hierarchical information base. So, in query response time, it is easy to process multidimensional modeling (for instance - compute total sales volume per product and store); operating with dimensions and hierarchies (for instance - roll-up: move up the hierarchy e.g. given total salaries per department, we can roll-up to get salaries per enterprise; drill-down: move down the hierarchy more fine-grained aggregation; pivoting: aggregate on selected dimensions usually 2 dims (cross-tabulation) );

comparisons (for instance - this period vs. last period - show me the sales per store for this year and compare it to that of the previous year to identify discrepancies); ranking and statistical profiles (for instance top N / bottom N - show me sales, profit and average call volume per day for my 10 most profitable salespeople);

custom consolidation (for instance - market segments, ad hoc groups - show me an abbreviated income statement by quarter for the last four quarters for my northeast region operations); etc.

Information Systems Conclusion The approach to build information complexes for information service of business accounting and decision making based on numbered information spaces which may support RTAP on the level of ODS Class I and this way to reduce the expenses for maintenance separate DW has been presented in the paper. This goal may be achieved using the FOI Archive Manager (ArM) and Multi-Domain Information Model (MDIM). An application of presented approach named Complex FOI was outlined.

Acknowledgments Author is indebted to Ilia Mitov and Krassimira Ivanova for support and collaboration. Due to theirs hard work the approach presented in this paper has been widely implemented in practice.

This work is a part of the project ITHEA XXI, partially financed by the Consortium FOI Bulgaria.

Bibliography [B-Jensen 2002] M.T. B-Jensen. High Tower Software's Tower View is the Odds-On Favorite of International Game Technology for Real-Time Data Management. Product Review published in DM Review Magazine July 2002 Issue. http://www.dmreview.com/article_sub.cfmarticleId=[Devlin and Murphy, 1988] B.A. Devlin and P.T. Murphy. An Architecture for a Business and Information System. IBM Systems Journal. Volume 27, No. 1, 1988. http://www.research.ibm.com/journal/sj/271/ibmsj2701G.pdf [Haisten, 2003] M. Haisten. The Real-Time Data Warehouse: The Next Stage in Data Warehouse Evolution http://www.damanconsulting.com/company/articles/dwrealtime.htm [Inmon, 1991] W.H. Inmon. Building the Data Warehouse, QED/Wiley, 1991.

[Inmon, 1995] W.H. Inmon. The Operational Data Store. InfoDB February 1995 http://www.evaltech.com/wpapers/ODS2.pdf [Markov 1984] . Markov. Multi-domain Access Method. // Proceedings of the International Conference on Computer Based Scientific Research. Plovdiv, 1984. pp. 558-563.

[Markov et al, 1993] K. Markov, K. Ivanova, I. Mitov, J. Ikonomov. Pyramidal Model of the Firm Information Activities. IJ ITA, 1993, Vol. 1, No. 2. (in Russian) [Markov et al, 1994] K. Markov, K. Ivanova, I. Mitov. Basic concepts and main information structures of the Complex FOI.

FOI-COMMERCE, Sofia, 1994. (in Bulgarian) [Markov et al, 1995a] K. Markov, K. Ivanova, I. Mitov. Automated service of the storehouses. FOI-COMMERCE, Sofia, 1995.

(in Bulgarian) [Markov et al, 1995b] K. Markov, K. Ivanova, I. Mitov. Automated service of the financial accounting using System "ANALYSE". FOI-COMMERCE, Sofia, 1995. (in Bulgarian) [Markov et al, 1996a] K. Markov, K. Ivanova, I. Mitov. Automated service of the accounting the staff and salaries FOI-COMMERCE, Sofia, 1996. (in Bulgarian) [Markov et al, 1996b] K. Markov, K. Ivanova, I. Mitov. Automated service of the accounting of the fixed assets.

FOI-COMMERCE, Sofia, 1996. (in Bulgarian) [Markov, 2004a] K. Markov. Multi-Domain Information Model. Proceedings of the ITC&P-2004 - International Conference "Information Technologies and Communications & Programming", Varna. FOI-COMMERCE, 2004, pp. 79-88. Int.

Journal Information Theories and Applications, 2004, Vol. 11, No. 4, pp. 303-[Markov, 2004b] K. Markov. Coordinate Based Physical Organization of Computer Representation of Information Spaces.

Proceedings of the Second International Conference Information Research, Applications and Education i.TECH 2004, Varna, Bulgaria. Sofia, FOI-COMMERCE 2004, .163-172 (in Bulgarian).

Pages:     | 1 |   ...   | 40 | 41 || 43 | 44 |   ...   | 54 |

2011 www.dissers.ru -

, .
, , , , 1-2 .