More on Schlumberger’s Delfi

Google blogger reveals inside workings of Schlumberger’s flagship ‘cognitive’ data platform.

Before we gave Schlumberger last month’s lead we did ask politely for more information on the breakthrough cloud-based infrastructure. None was forthcoming. However, Schlumberger’s information-retentive guardians of the truth forgot to tell partner-in-crime Google who’s SVP Urs Hölzle has been blogging away regardless. Hölzle, reprising his address to the private, clients-only Schlumberger Forum in Paris last month, provided more on Delfi’s innards. So, according to Google, the Delfi E&P data lake is based on Google BigQuery (data warehouse), Cloud Spanner (RDBMS) and Cloud Datastore (NoSQL) ‘with more than 100 million data items, some 30TB of petrotechnical data.’ According to Hölzle, Schlumberger’s petrotechnical flagship, Petrel, and the Intersect simulator are running in the Google cloud, ‘integrated into Delfi.’ WesternGeco’s Omega geophysical data processing is ‘running at a scale not possible in traditional data center environments’ thanks to Google cloud-based Nvidia GPUs and ‘custom machine types’ giving a compute capacity of ‘over 35 petaflops* and 10PB of storage.’ Other novel tools include TensorFlow, open source artificial intelligence, used for log QC and interpretation and also for 3D seismic interpretation.

Hölzle reports that Schlumberger has deployed the Apigee API management platform (acquired by Google in 2016) to provide ‘openness and extensibility’ allowing clients and partners to add their intellectual property and workflows into Delfi. Read Hölzle’s blog here

Chatting with some Schlumberger people at the SEG it was unclear how much, if any, of this laundry list of Google’s software is fully deployed in Delfi.

* Although SLB’s Paal Kibsgaard reported that Schlumberger had a 27PF capacity way back in 2013.

Click here to comment on this article

Click here to view this article in context on a desktop

© Oil IT Journal - all rights reserved.