In this part we will try to expose some major faults related to practical dealing with voluminous geological, geophysical and production data. We will concern, in particular, problems the customers meet when they employ the software (data banks and interpretation systems) built on relational DBMS foundation. In addition, we will point to the semantic object-oriented database use as the possible way of involved problems' solution.

Nowadays the relational DBMS's are rather commonly used for geological, geophysical and production data manipulation. ORACLE occupies the leading position. It is used as a platform to design many software complexes, such as, for example, Finder (Schlumberger), PDS (IBM), PetroVision (CGG) and so on. These and other packages also using relational DBMS's are actively introduced now not only for particular tasks solving but for global data banks creation at the level of an enterprise, corporation and even territorial-administrative district as well.
The relational DBMS's domination (ORACLE in the first place), their marketing promotion by the Western companies have resulted in firm conviction that the table data model is the most efficient and nearly the only acceptable means to operate information. Meanwhile there are quite a few examples of successful use of above- mentioned software packages. Their adaptation to meet peculiar customer conditions takes, as a rule, the years, and hardware capacity requirements grow along with soaring data volume.
An explanation of this contradictory situation seems to be as follows. Software packages developed on non-relational databases are much lesser part of the software market, and their developers do not possess appropriate marketing capacities. In connection with this the relational approach is not challenged in practice by any competitor. Therefore customers are deprived of possibility to be convinced that some serious problems they meet with are not inevitable in principle but result from inefficient approach to software design. Let us briefly specify most evident faults of the relational DBMS's and problems they engender.
First thing one should note is that software packages utilizing the relational model (usually ORACLE) make great demands on the hardware. Now use of workstations for geological, geophysical and production data processing becomes quite common. They are believed to be the only means to provide for currently necessary productivity of data management process. There is not need to remind that the workstations are rather expensive (as compared with the PC, for example) hardware. They not only cost quite a lot but also require significant expenses and special equipment for maintenance. To install the workstations is simply impossible or extremely unprofitable in those industry departments where appropriate technical conditions are absent (for example, in branches). High price and "capriciousness" of the workstations become serious (may be crucial) obstacles to their widespread use. One part of potential customers can not afford them due to lack of money while another part can not exploit them on technical reasons as well (working in the field conditions specialists are meant, for example).
Above-enumerated intrinsic to the workstations restrictions are not usually considered as the faults. They are believed to be stipulated by the very reality. However having known alternative methods of information tasks solutions (they will be told about beneath) we have the right to assert that demand for the workstations use is stipulated only by inefficient data model exploited software packages are based on. Another serious fault we connect with the relational approach is a poor adaptability of appropriate software tools to individual customer conditions. Relational databases deal with strictly specified parameters sets, logical integrity of stored information being ensured at the database query level (application level). Therefore, any parameters specification change results in correction of not only database structure but involved application software as well. Practice proves that such a database adaptation may take the years. Meanwhile the industry enterprises' and corporations' activity scope as well as employed technologies character are permanently changing. New parameters categories periodically emerge. They need new positions to be allocated at the data bank structure. The relational DBMS's use experience proves that they cope with this problem rather badly. Table-located data volume growth as well as new tables addition rather drastically reduce database queries efficiency. Upon such a difficulty having occurred the customers are usually proposed to purchase new, more sophisticated hardware and, later, suitable software too. As a rule, the customers consider those measures as quite necessary. In such a manner the lion's share of funds (about 80%) is usually spent on senseless, unproductive tasks solving. One can not miss that the situation is extremely beneficial for numerous manufacturers of workstation hardware and software.
To overcome involved problem the relational approach supporters suggest strict standardization of both the data categories and data representation means. Actively developed now standard POSC is an example of such an approach. One of the problems it is intended to cope with is to foresee all data types to be probably dealt with in the future. Taking into account high-grade data volatility and substantial data processing methods difference at various enterprises one can easily conclude that the attempt seems to be deprived of any logic.
Information logical integrity maintenance task (which is solved at the level of relational database queries) compels the POSC standard developers to stipulate strict binding of various data types. It is expressed in the rule that one category data must be inserted into database or deleted only together with some other category data bound with the former. Therefore, to comply with this rule inevitably compels the customer to adjust his pertinent tasks to the software at his disposal.

As it was many times noted unfavourable for a customer situation results from the inefficient database management system used as a foundation for proposed technology solutions. We may assert that nowadays there is a real competitor with the relational approach. Semantic object-oriented DBMS forming foundation of new data management approach is almost completely deprived of above-mentioned defects. The semantic DBMS important feature is an orientation to the objects which are information models of objects, events and phenomena of the real world or, to be more exact, of involved application field. In such a database any parameter value is allocated as a single instance, this parameter presence in various objects description being realized by means of special references mechanism. This ensures database information minimal redundancy and, moreover, significantly reduces amount of possible errors.
Every parameter being considered as a potential object results in perfectly painless data sets upgrade.
Another important benefit is that redundancy absence and high-grade data logical integrity provide noticeable reduction of requirements to the computer resources. Enumerated benefits of the semantic object-oriented data model are confirmed by profound but not well-known yet practice. Long-term use of the information analysis system CTC-EXPERT under conditions of enterprises in Russia and the CIS countries has proved that semantic DBMS CTC forming foundation of the fore- mentioned system is an efficient data management tool deprived of all above- enumerated faults. In the first place it is important to note that DBMS CTC operates on the IBM PC platform. Computer equipped with Pentium processor and 2-Gbyte hard disk is quite enough to realize storage and full-quality retrieval of exploration and production data related to vast (more than ten thousand wells) oilfields. Such a configuration provides roughly the same data retrieval efficiency which is attained at the workstations using ORACLE. Benefits of the PC exploitation are evident: firstly, much lesser cost of both the hardware and its maintenance and, secondly, computer (computer technology) availability at virtually any industry sector including primary departments and oilfield plots.
None computer technology is of any practical importance for an enterprise until it is adapted to individual conditions. When geological and geophysical data bank is meant, such an adaptation must take into account not only involved parameters specification peculiarities and processing techniques, but also the ways of the data import and export. The semantic DBMS allows us to react adequately to both of the factors. Flexible DBMS makes it possible to accomplish its quick adaptation to the customer conditions and to organize effective new technologies maintenance and support. The semantic model is so universal that it permits us to consider data of other structure (table, for example) as a subset of proprietary data. This fact is of great significance for adaptation task because it allows us to include data files of various formats to the semantic DBMS structure in a natural way.