Speaking at the 2009 PNEC Data Integration Conference (report on pages 6&7) in Houston last month, Chevron Fellow and Master of Upstream Architecture Jim Crompton asked, ‘Why are we still challenged by data management?’ Crompton noted that many issues discussed at earlier PNECs remain. Some folks believe things were better 30 years ago before we went digital. The good news is that management is now concerned about data management but the bad news is that management usually turns to IT and says ‘fix it.’ This is not the way forward.
Meanwhile, the need for current data is growing. Users want data now and they don’t want to spend time making sure it’s right. This is especially true in the context of production and real time data. A multi disciplinary approach and asset-level partnership is needed. To date there have been some IM successes in restricted domains and geographies. But things fall apart when a production engineer needs data from the drilling system. For work in a collaborative environment such as an RTOC, many IM problems need to be addressed. Crompton showed a complex spider diagram of multiple systems of reference (SoR). Chevron has identified around 40 SoRs needed to run the business. So far only around 20 SoRs are under control—the rest are on spreadsheets or otherwise not managed! Chevron’s information architecture (IA) is an overarching approach to IM including quality improvement, governance across the information value chain. ‘You can build a portal in 6 weeks but it might take 6 months to figure out and solve the underlying data.’ The cornerstone of IA, the Keystone master data management (MDM) server, provides the ability to interoperate across key information objects. Here, ‘we are talking about the kind of data that most users don’t want to know about.’ In other words ‘referential standards’ rather than ‘how much is my well producing’. It is hard to get funds to attack this sort of problem and even harder to define cross discipline semantics in a way that is satisfactory to all. But without a common semantic there are ‘too many good meanings.’ Problems go from hard to ‘really intricate’ and the question is, ‘when do you stop analyzing?’ Keystone includes safety issues. ‘Bad data can hurt people. It is not just dollars and cents and the ROI of MDM may not be obvious to all. What we are attempting to do involves well, legal, land and more—all the semantics of our industry.’
Last December we reported from the Chevron-hosted Semantic Technology in Oil and Gas event where there was talk of using semantics to manage enterprise master data. Crompton told Oil IT Journal however, ‘While Chevron has several research projects around the semantic web, the Keystone project is not planning on using this technology for our Upstream Master Data Management approach at this time.’
© Oil IT Journal - all rights reserved.