Shell’s Reference Architecture

Shell’s global data architecture V1.0 rolled out mid 2010. Under the hood are IBM’s System Architect and Metastorm’s ProVision. Inspiration came from the DAMA DM-BOK and The Open Group’s TOGAF.

At the IRM-UK data management conference in London late last year, Andrew Schultze1 described Shell’s journey along the road to a global data architecture (GDA) strategy. Shell has a long history of data modeling—as witnessed by our review of Matthew West’s new book on page 3. But before GDA, the behemoth’s (100,000 employees and $13bn income in 2009) decentralized business meant that its data architecture was divvied up between multiple units. The new initiative—described as an ‘ongoing journey’— set out to provide a terminological framework, an enterprise data catalog that will allow subsequent ‘deep dive’ data modeling into specific areas.

The top down approach began with a management decision that an enterprise data strategy was needed spanning the whole upstream-downstream spectrum. Shell’s initiative stemmed from a business ‘appetite’ for improved data quality along with the realization (from its PDO unit) that data architecture can contribute to better data quality. Shell has now developed an enterprise data catalogue (EDC) that graphically represents its data at a high level. The color-coded EDC shows which business leads for a subject area, where data quality dashboards exist and where ‘deep dives’ have been executed or are planned. The EDC lets Shell scope projects and identify overlaps—avoiding ambiguity with a single terminology. The EDC provides Shell with thematic direction, ‘big rules’ that programs and projects can follow. Users are encouraged to ‘be pragmatic, boiling the ocean will not be accepted.’

Tools used included IBM System Architect and Metastorm’s ProVision, a toolset for business and information architecture that claims to deliver a strategy-driven business and IT architecture (ProVision is also used by Hess, Talisman and RomPetrol). Part-way through the project Shell decided also to leverage the DAMA Data Management Book of Knowledge (reviewed in Oil IT Journal in June 2010) as a source for terminology.

The EDC currently lacks a lot of stuff and is to be expanded, ‘borrowing’ from industry standards such as ISO 15926, ProdML and others. V1.0 of the EDC was released in April 2010.

Shell is now working on its data management processes and on master data management—starting with upstream data from its onshore US gas business, turning to client and customer data in 2011. Product, pricing materials are also in the GDA—as are records management and ‘e-discovery.’ Shell’s acronymic bonanza is topped-off with a project delivery framework (PDF), hooked into the GDA.

1 The presentation was made by Lars Gåseby and followed-on from Shell’s 2009 IRM-UK paper where Johan Krebbers described how Shell’s enterprise architecture has been constructed leveraging The Open Group’s ‘TOGAF’ framework—

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.