Given a large, integrated oil and gas company which wants to completely renew its IT/IM, what do you propose?
Bayne—Minimize complexity as follows. Minimize integration complexity via Seabed as a common logical data model. Minimize tool complexity with common applications and data-stores and finally reduce the demands on end-users by enabling workflows like our Petrel plug-ins that deploy IM workflows natively.
Bouffard—End users are looking to new ways of working—these may not be completely defined. But we do have some technology directions to offer. Streaming RT data is becoming part of daily life. Access to production data—WITSML, PRODML. We see convergence of structured and unstructured data.
Bayne—You will always need structured databases but there is huge growth in emails etc. Here we defer to Infosys for help in managing this.
Bouffard—IM may not always be apparent. For instance in Shell’s HDP*, interpretation and data access all goes through the Portal. The future will see interpreters doing more and more in less and less time—and data management will have to be done in the background.
Why do we even bother building projects? Why can’t you offer interpreters windows onto their data à la MapQuest?**
Bayne—Our Seabed paradigm is a step in this direction.
Bouffard—I agree that today’s technology is not up to this. We expect to be able to offer such facility through web services and data replication.
We hear a lot about web services (WS). Where is SIS in WS, with self describing data, tagging, data validation?
Bayne—ProSource and Petrel DM leverage web services today. Our integration engine as deployed in ProSource Results leverages a services-oriented architecture.
Can workflows be captured in XML, edited programmatically and re-run?
Petrel does not support the ability to export the workflows to XML currently. SIS is experimenting with Microsoft Windows Workflow Foundation to enable this kind of functionality.***
Talking of ‘openness’ in the context of Petrel is something of a paradox. Petrel is a facet of SIS’ ‘wall to wall Microsoft’ approach to technology. A far remove from what is generally understood as the Open Source movement!
Bouffard—That’s not quite how we see it. At one level, the Ocean API provides openness. But users won’t get access to all of our IP as you would expect from an Open Source environment. But Ocean does allow you to add your own functionality to applications like Petrel. We are completely genuine on this openness issue. Publishing Seabed has been a major event for us. This data model represents tens of man-years for our data modelers. Offering this up to the market is a bold statement and we hope that the industry as a whole will gain from this initiative.
We had a look at the Seabed model—it is a text-document, are you planning to release an implementable DDL version of the data model like PPDM does?
Bouffard—PPDM restricts its DDL to members. SIS ‘only’ publishes the Seabed data model—but makes it available to all. Indeed, we do perceive the physical implementation of the data model as a differentiator. Your readers might like to check the model on www.slb.com/content/services/software/opensystems/index.asp by clicking on the Seabed Data Model link.
* Hydrocarbon Development Planning
** An idea we developed in our September 2006 editorial.
*** This answer was supplied offline.
© Oil IT Journal - all rights reserved.