When we started out back in 1996, as Petroleum Data Manager, data management was, on the one hand, the topic du jour. On the other hand, there was a feeling that, mirroring Fukuyama’s ‘end of history’, data management was a ‘done deal’. Depending on your persuasion, either what was then POSC (now Energistics) or PPDM* had fixed, or would shortly fix, all of what were then the pressing issues of the day**. That is to say, the pressing issues of the data world that have remained with us ever since; awkward data formats, idiosyncratic data stores and multiple issues of interoperability. The very issues which OSDU, the Open subsurface data universe, has set out to solve.
This issue extends our coverage of OSDU with a technology update, a report from The Open Group’s Amsterdam meet (with also an update on the other TOG oil and gas initiative, O-PAS, the open process automation standard) and interviews with OSDU champion, Shell’s Johan Krebbers and TOG CEO Steve Nunn.
Will OSDU fix interoperability? To believe so, you have to believe that ‘this time is different’, which is always a big ask! On the plus side, OSDU has momentum. But momentum is not enough. There is a tendency amongst major operators to spin-out putative standards work at a certain stage of development in the hope that they will see take up or sustainable further development. Such in-house work is generally presented as non-core business. Over the years we have seen the following standards spun-out (mainly from Shell) into the public domain: ISO TC/184/15926 with its Express data modeling, Business Objects with the spin-out of Open Spirit, revamped ISO 15926 with added ‘Semantic Web’ technology and lately Cfihos and its standardized Excel engineering spreadsheets***. These all have had a mixed reception in the real world for a variety of reasons which Oil IT Journal has covered extensively over the years. I think it is fair to say that they have all fallen short of their original intent, although to be fair, it is too early to pronounce on Cfihos. Today Shell is spinning out its in-house SDU infrastructure into OSDU. Will this time be different?
I really don’t know, but one interesting facet of OSDU is its focus on open source software. Johan Krebbers told us that ‘We can’t have any proprietary stuff on the platform’. Which begs the question, why are so many vendors of proprietary software interested in the initiative? Have they all undergone a damascene conversion to free? On the side of the oils, a similar conversion is required in regard of open source. Open source (apart from Linux in HPC) has had an unenthusiastic reception in oil and gas to date. We have tracked the open source movement in the upstream out of interest and I confess, from something of an activist stance, dating back to the days when it was IBM vs. Unix, or Microsoft vs. the ‘cancer’ of Linux. There was general disdain, and even fake news, with whispered warnings of legal repercussions if you were caught using open source software. Industry events targeting open source, notably chez the EAGE, failed to get much traction.
Today, the GAFAs have changed the nature of the open source game and even Microsoft has ‘embraced’ open source****. What oils need to do now is get into the open source culture. That means a shift away from exclusive in-house work, from closed R&D joint ventures, and into writing and financing a code base for the public good. The activist in me adds that the code should not be ‘open source .NET’, which would make it exclusively Windows. Neither should it be too finely-tuned to a particular vendor’s data universe. This makes for quite a tricky balancing act. But caveats apart, OSDU is the most interesting development in the upstream software world since the heydays of POSC’s Epicentre.
* At the time this was the Public Petroleum Data Management association. In 2008 PPDM changed its name to the Professional Petroleum Data Management association.
** Notably with Project Discovery, which in 1996, set out to merge the Epicentre and PPDM data models.
*** Parallel ‘open’ offerings from the major vendors have also emerged but these, from Landmark’s OpenWorks to Schlumberger’s OpenDES, are too tightly-coupled with their proprietary infrastructures to be considered truly ‘open’.
**** Although this could be interpreted as continuation of its ‘embrace-extend-extinguish’ strategy of the 2000s.
My apologies if you have heard this story before. A mysterious traveler arrives in a mediaeval village in the midst of a famine. The traveler promises to feed the starving villagers with a magic stone he has in his knapsack but first he needs a soup cauldron and some boiling water. This duly arrives and he pops his stone into the pot saying, ‘That’s it. The soup is cooking now and will be ready soon. But, you know, it will be nicer if any of you have some vegetables to add in for extra taste’. The headman disappears briefly and comes back with a large parsnip he has been hoarding and pops it in. One by one, the other villagers do likewise, until the caldron is full of boiling veggies. A few minutes later, the villagers tuck into a delicious broth and thank the traveler for doing his magic.
I’m not sure that I did the stone soup story justice. Wikipedia has a few interesting versions. I like the Hungarian one which has it that, after the meal, the traveler sells the stone to the villagers. Which fits my point better. In the IT world, the stone is the ‘next big thing’. Over the 25 years or so that Oil IT Journal has been publishing, the next big things have included the relational data model, CORBA, business objects, the semantic web, and more recently, cloud computing and ... Kubernetes. The vegetables are your domain knowledge and your data.
© Oil IT Journal - all rights reserved.