OITJ Interview—David Gorsuch

Ocean project manager explains how Schlumberger’s ‘open proprietary’ development environment will offer third party developers access to the new .NET-based products, including Petrel.

OITJ—What exactly is Ocean?

D.G.—Ocean is an open proprietary development environment, designed to foster innovation and creativity. Just as Microsoft’s Visual Studio lets developers create software that runs on Windows, Ocean will let third party developers give their tools the ‘look and feel’ of SIS* mainstream software like Petrel, Peep, Osprey and Avocet. There is a wrinkle to this approach though. One development environment won’t hack it for all products. For seismic to simulation (S2S), the Petrel 3D canvas is ideal. For drilling, value and risk (DVR) the Petrel interface is not appropriate—engineering tools need a more flexible look and feel. The Ocean development environment supports both S2S and DVR in two distinct products. Common to both is the Ocean Core and a set of services for coordinate reference systems (CRS), units of measure etc. There are few things that irritate our customers more than seeing the same well in two places because of different CRS implementations.

OITJ—When will developers get their hands on Ocean?

D.G.—The first commercial release will be towards the end of 2005. At that time, developers will get access to the Petrel infrastructure, the 3D canvas and rendering engine. We will also expose the Petrel models and some seismic data types. By the end of 2006 the drilling world will be included and data access will include more seismic data types.

OITJ—Ocean has been touted as a .NET development environment. Does that mean that it will be ‘web services’ based—or just a recompile of a ‘monolithic’ Windows 32 bit MFC** application with the latest version of Visual Studio?

D.G.—Initially it will just be a recompile. When .NET was announced, Microsoft completely blew it by emphasizing its ‘web services’ aspects. In reality, .NET is a development environment like MFC. Web services are still taking shape. But by going with Visual Studio we hope to position ourselves to take advantage of web services and Services oriented architecture when they’re ready.

OITJ—So you’re abandoning Linux?

D.G.—No, we are not turning our backs on Linux which we see as having an enduring position in the cluster world—especially for compute intensive applications like Voxel Vision, Eclipse and Decide!

OITJ—Not GeoFrame?

D.G.—We see a long and prosperous future for GeoFrame, but we expect clients to transition to the new environment over time. We are not going to hustle folks. GeoFrame on Linux is very performant.

OITJ—A move from Linux (or Unix) to Windows is predicated on 64 bit computing. When will Petrel be 64 bit?

D.G.—I wouldn’t say ‘predicated,’ but Petrel will be 64 bit in the near future. We are re-engineering the Petrel interface to be ‘64 bit ready’ and of course we are waiting on Visual Studio 2005, which will be 64-bit enabled. Unfortunately, VS 2005 has been delayed, so all I can say is that Petrel and Ocean will be 64 bit ‘real soon now’!

OITJ—So the plan is to recompile Petrel on the VS 2005 64 bit when it is available?

D.G.—Not exactly. By year end 2005 there will be a new version of Petrel which will migrate the Petrel GUI to the .NET WinForms. There is a lot of MFC stuff in Petrel. The key is to get the user interface in .NET and then start testing on 64 bit.

OITJ—What is ‘open proprietary’?

D.G.—SIS will own the source code and development environment. We will publish the API with user guides.

OITJ—You said ‘anyone’ can buy it. Would you sell Ocean to Landmark?

D.G.—Yes, but I’ll be buying a lot of people drinks when that happens.

OITJ—So it will be different from the old days when the GeoFrame dev kit was not exactly ‘open’.

D.G.—Yes it will be different. I am adamant about a separate identity for Ocean as a product with attention to documentation, presentation etc.

OITJ—Where does this leave OpenSpirit?

D.G.—OpenSpirit (OS) is SIS’ provider of data access middleware. Customers will use OS to access data in legacy data stores. We are actually working with the OS folks to move their data access middleware forward. They are challenged by the rigidity of the data footprint and we are working to move data access to a meta data driven approach, a key piece of the puzzle. OS needs to be more flexible.

OITJ—What is ‘metadata driven’?

D.G.—You’ll be able to create metadata to define mappings between standard data and data store specifics. This will move from a ‘canonical’, hard wired model to a flexible system, making it possible for OS to provide third parties with the capability of writing their own adapters. This will grow the OS footprint. Also if you want to include a legacy data source, you can use your own metadata to get at it—without having to wait on the OS footprint to ‘evolve’.

OITJ—What was in the 2004 Petrel API?

D.G.—The Petrel API exposed a subset of the Petrel data model letting developers create Petrel plug-ins—a first step. The 2004 API allowed access to four Petrel data types and let users create processes that appear in the Petrel process manager and become part of the Petrel workflow. This API has a algorithm-focus. You can build you own facies modeling algorithm, but you can’t access Petrel infrastructure or create new data types.

OITJ—Will Ocean’s support for geodetics, units of measure extend to UWI*** and other reference data? In short is this the answer to Steve Comstock’s plea for better Petrel data management?

D.G.—There is no ‘policing’ of metadata, UWI, CRS etc. built into Ocean. We see this a separate IM challenge. Petrel IM will leverage the new common Seabed data store. We, and the users, like Petrel the way it is and we don’t want to mess it up. So we are keeping Petrel ‘clean’ and letting users store project files in the database rather than on the ‘C’ drive. In the future you’ll be able to ‘decompose’ a Petrel project file into a Seabed data store, keeping the stuff that people want to share. Ultimately such files will be decomposed completely and the Petrel Project File will be a pass-through cache to data stored in Seabed.

OITJ—Was last month’s endorsement from Microsoft’s Steve Ballmer any more than product placement?

D.G.—SIS is now one of Microsoft’s strategic partners. But that does not mean that we are going to have much influence on the 440 million Office users! We do expect to see opportunities for collaboration in certain areas, like high performance computing where Microsoft is faced with Linux/Unix dominance. Microsoft creates a load of ‘stuff,’ we want to know about it as it happens so we can help early adopters.

* Schlumberger Information Solutions.

** Microsoft Foundation Class—the Windows API.

*** Universal well identifier.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.