The modeling malaise, in E&P data management (July 1996)

PDM offers a brief retrospective of the data models competing for market share in E&P computing.

First of all, contrary to what one might be led to believe by the current talk of data modeling coming from both vendors and standards organizations, data management is not synonymous with data modeling. Nonetheless, data modeling is an important building brick in data management, and current developments, particularly in the relationship between the Petrotechnical Open Software Corporation (POSC) and the Public Petroleum Data Model (PPDM) make this an extremely interesting starting point. The reason for all the attention given to data models today is that they are a way out of the interoperability impasse described in the previous section. Most applications today run on pretty much the same sort of hardware. They run on the same Unix operating system and even frequently use the same database management system. The next step is therefore to use the same underlying data model.


If all applications know that, say, well deviation data is contained in a database table called WELL-DEV-DATA whose structure is known beforehand then there only need be one copy of this table available to all applications, greatly simplifying data entry, access and maintenance. Now the foregoing is a simplified account of the situation prevailing about seven or eight years ago when a number of different organizations began to address the problem.

Other players

In Canada, PPDM started out with a very straightforward mission, to build a data model running on the mainstream technology of relational data bases such as DB2, ORACLE and SYBASE, and utilizing a standard programming language developed in the 1970s by IBM called Structured Query Language (SQL). More or less simultaneously, a group of major oil companies in the US and Europe clubbed together to establish POSC which although sharing the same goals as PPDM, had a more ambitions view of the problem, and who were prepared to entertain more forward looking solutions. Others also entered the data model fray, Schlumberger who put their Geoshare technology into the public domain, and Petroconsultants came out with IRIS21, a data model closely coupled with Petroconsultants "raw material", land lease and well information. Meanwhile, EXXON and IBM with the Mercury product, had been modeling data for some time before POSC and PPDM came on the scene.


Now you will already have spotted the problem with this excess of enthusiasm for data modeling. Too many data models! In fact it can be broadly stated that the data modeling effort rather than reducing the workload of reformatting data from different applications has actually added to this by introducing a new dimension to the solution, that of "mapping" data from one data model to another. There, very briefly, is where we are now. We are still pumping data from one format to another via for instance SEGY. We are also playing a new game, that of finding common ground between different data models. The fruit is still on the tree, we haven't found the ladder yet. The Discovery project (see separate article) describes a recent attempt to find common ground between PPDM and POSC.

Click here to comment on this article

If your browser does not work with the MailTo button, send mail to with PDM_V_2.0_199607_7 as the subject.

© Oil IT Journal - all rights reserved.