 This session will explore the business challenges that oil and gas... Okay, yeah, still morning, good morning. So we're going to talk about the subsurface space for people who don't know, well I don't know who's holding the audience, for the people who don't know what subsurface of course is, is the first beginning of any oil and gas company, so it's the exploration space where you try to go out there shooting seismic, it's the development space where once you have hopefully found something in what the developed that field and it's the wells space where you start really getting your production wells in space. That together we call subsurface space and we're going to really looking at why we're doing this of course there's a lot of change happening in our industry because of course the whole trend to know towards new energy sources at the same time we still have oil and gas, a special gas for many years to come to really support our demand for energy around the world in a growing population and of course which is a bigger impact is the move to digitalization and this stage is all about data, I mean if you talk about this stage you really talk about data, how do you exploit data, that means that's the whole story around digitalization. That means we had to start working in a very different way and that's what this is about, how do we get, how do we make that happen in our industry. So first I'm going to talk about the why, scope, the members, current status and plans for the future. So think about the subsurface space, if you look at that whole space of subsurface you've got many disciplines and each of these disciplines have their own applications and really the data is owned by these applications. So the problem you have you've got many silos so if you look all the way from from exploration to the wells environment you might have several silos of data and of course it's very hard to look at the end to end picture. Therefore the third bullet point how do you see a view if you try to drill a well, can you go all the way back to exploration what really happened. That lineage is very difficult to do today because all that data is linked to the very disciplines, you've got lots of silos out there. So if you really want to be ready for going from a physics world to a machine driven world where you really have to need to bring all the data together and then use that data for for fault finding or other things you want to do you have to bring that data together into a single data platform and move away from those silos for the application drive. So you move to an application driven world to a data driven world and that's what it's all about. We, I mean it's everybody that is not unique to the subsurface space, it's also not unique within itself, it's true for the whole of Shell of course, but here I'm talking about the subsurface space. So you have to move to a world where the data is owned by application, the data becomes one large data platform. You talk about fairly large environments, so in a typical integrate all the gas company like ourselves you talk about maybe a hundred petabytes which goes into a single data platform here. If you look at all the data you're shooting all the size when you're shooting all the other files are out here you become too about that size if you bring it all together. So how are we going about this? So we said okay we of course can do this all on our own and try to make this world so really and we started this development called the SDU, the subsurface data universe. So we started the development about a year, yeah nine months ago, started developing a data platform showed in the moment with the application around it and we said well of course we can do it as Shell on our own but does that make sense? It's far more important now this market to really try to go to an open source based environment for the speed of change in open sources many many times faster than any prior provider environment. So we said we looked at it, we looked at it and we said in March we start bringing together other operators say hey can we make this an indices solution? What we're talking about of course you talk about let me first show this picture so this is really the elements we're talking about at the bottom you see of course you infrastructure platform this is all a public cloud environment we only talk about public cloud no no on-premise unless we need to do like on the left because we have we operate in countries where we cannot export the data to other parts of the world. We look at the bell time you see they're hosting environment pretty straightforward the crux of the system is your data platform yeah that's where you store all of your 100 petabyte of data of course there's lots of different data source out there it's not all structured data a lot of it is unstructured data binary data is a whole raft of different data sources therefore your data ingestion source is so important because the data ingestion services really extract metadata if you put if you have a large data platform you put millions and millions of files in the environment if you have no metadata you create yourself a big problem of course because all you have lots of data we have no information about that data so therefore the data ingestion is really about extracting metadata of each file you're loading that environment yeah that is an and and a wells file a drilling file we don't really care you extract metadata and the metadata you store in a graph database in a data platform so data ingestion and data platform with the crux of this hell environment on top of that you have your business application your command applications your all microservice driven environment so your command applications your individual applications your orchestration services of course the html5 based interface all browser based environments then you've got your information security but if you store lots amounts of data you also have to think about your security model how to how to protect that data of course the data will be accessed not by shell people many other parts of the industry as well of course support environment over there that's really if you talk about the subsurface data universe that's the pixie we're talking about on the left as I mentioned we operate in countries like many oil and gas companies where we're not allowed to export data and bring to a cloud environment therefore you also have your in-country solution over there so if you talk about what we try to bring into the open source environment this this environment so if we word word u.s. companies differentiate or compete is in the application space yeah so application space is really where you want we'll compete between each sort of but also with other software companies now there the rest of environment the data platform the hosting security is really very much a common based environment so what we try to do is create a reference architecture for this so people can use as the base we've talked about what the covers it covers the exploration space the development space and the well space and all the data you can think about there and that's a whole vast array of data and also increase the real-time data because increasingly we have sensors in our fields collecting real-time data over the lifetime of that asset and all the service we talked about it your data your storage your compute your application platform etc is all part of this scope we're talking about and of course only public cloud based so what we what we what we try to do now is is really to say okay how can I make this how can we make this an industry solution not just a shelf solution as a man so we started that exercise in March of this year we had together about eight eight operators and i'll come to the names in a moment and we now pursue this under the umbrella of the open group to start a new forum to really try this as an industry solution not just for the eight companies out there but also for any all a gas company or all software companies supplying to this market everybody is really we can become a member of this of this new to be established forum so what are the two things we try to see a common data platform for all exploration development and wealth data yeah and a common platform for development deployment and running of applications so make sure they have a single platform with well-defined apis on top of which anybody can develop any type of application and will fit all that platform whether a platform runs at Shell or BP or at Exxon or at Chefron or Devon or whatever it is so we create a market out there which goes beyond a single company that's the whole purpose over here yeah and then the last point over there of course make it that as part of an open source environment the bottom line is why we're doing this is creating a market out there where people can develop applications against and really compete where we should compete and really use open source where we can use open source so the lay of apis of the data platform is essential for people to start developing any type of application on top of that so the data as I mentioned is it's really the single data platform as I mentioned the data ingestion services all part of the standard environment on top of course a hosting platform and then the rest application on top so these are the parties at the moment working this so these are people the companies kicking this off together with the open group under umbrella of the open group and we come together tomorrow over here to have a another face-to-face meeting to talk about the release the first release etc and the formal name is I think was on the first slide but I'm not sure I got on the slide but the formal name is now the open subsurface data universe forum there's the formal name will be called and in addition to the other ones over here but it's then the open subsurface data universe forum so the focus is as I mentioned at the subsurface space anything in that space so where are we current status so we had a number of meetings under or species of the open group to bring together okay are we interested do want to do this yes we want we want to do this we are finalizing the charter so the charter still in draft we hope to finalize it by next week the July 31st and then printed up goes in the final charter and now we're starting to talk about okay what will the first release look like to be issued I would say late this year early next year and how what will be issued what are the standards the api is the reference model we're going to issue etc so the focus now starts moving to a content they were based on a work already done in Shell but also the work been done in other companies because as a means we start is working Shell about nine months ago so that's the kind of material we also will use in this outfit to make sure we have a kickstart in this space and it's really the last slide this really want to want to want to talk about so it to repeat it really costs subsurface space it's really making sure they're being this whole world to an open source environment why do we need to have much faster speed of development and of course we're moving from a silo based approach to a data platform based approach because of the move to and this is a safe environment driven by machine learning etc so that's what we're trying to achieve and that's what we've been kicked off more or less today so thank you much for your attention Johan thank you very much for coming along and explaining what's behind this and please have a have a seat we'll move a microphone or something so as as Johan has said we we have the effectively we have a new forum in the open group which is which is great I'm always happy to announce those the organizations involved so far in this have have been moved very quickly what we've seen and correct me if I'm wrong but what we've seen is that Shell had a a business problem to solve and very very quickly identified that their problem was probably not unique to Shell but probably to other operators the other operators all that so far involved have all resonated also with the with the problem and the idea of pooling resources to try and solve this problem create a common a common standard around this a reference architecture is is what the open group does brings people together to do things like that so it's a it's a it's a nice fit and as of today anyone interested in in joining the forum and in part in participating is is able to do so welcome to do so still in the formation stage as as Johan has said so there are some some things to tie up hoping to get a charter by the end of the month but it's it's going to be a new and exciting forum so first question to ask you is is we're starting with operators what other type of organizations do you think may be interested in in membership of here in on the supply side for example yeah so if you look at the operators you got you got all the major iocs there but of course there are many other operators out there if you look at the national all companies so we're really trying to create a market out there for other parties to develop software or hardware mainly software for this platform and it can be software supplies in this market that can be university it can be academia it can be anybody who wants wants a broader platform because often there's a problem in the gas industry we have maybe too many platforms and often they are very small and if you really want to survive over you have to really make that platform much larger and much more popular so it's important that you create demand and create a market out in principle these eight or nine companies starting doing that and and of course when you start creating the people developing software for that you get the chicken and egg story more software developed but more people join more people join more software to be developed that's all about the platforms all about it's if you want to survive in today's market you need to have popular platforms you can see all around in industries but it's all about creating a popular platform in subsurface space there'll be multiple applications for the application can be very small it can be very large it's really market service different environment we'll have clear frameworks how people can develop for this platform but really our aim is to create a market out there for people to develop and that can be any software company are they active today or not active today and should be of interest to cloud providers as well as part of it so a number of questions have come in while you've been speaking and people are hearing about this for the first time by and large in this audience so will the objective be to create a google earth type functionality for what's under the ground google earth also includes the application layer our first focus also with google will be the first focus is purely creating that that data environment so to keep it bring that data together in the common in the platform and of course on top of that you can have google earth as an application I mean search of course is a very I mean that's why I mentioned before the metadata story because with metadata I can enhance my graph environment I can do my search because search in that in essence is usually metadata story yeah so in principle it will google earth can be used but of course let's be clear not all the data values every operator will be in one environment there will still be a shell implementation of this environment an external implementation or not what you call it and chevron implementation of this environment it then does not bring all the data together every operator into a single last database is still a shell and today shall implement this standard so will the other parts on the on the on the thing but in each of those environments yes you will well the key over here that's why I mentioned the silo what's key over here thanks to the metadata I can see which we can't do today if we do something let's say in Australia I want to know we have a do we have a similar structure in other parts of the world and how would it look like today that's almost impossible to do because of the silos we have so if I want to know in this environment see okay I have a problem in Australia I want to have a certain structure say why else in the world do we have that in this environment can do that I could principle look anywhere in the world what they do have available what structures do have available and how what can I learn from that today that's not something which is really available okay great so I think you I think you answered this question but it in the negative which is is the plan to have a data platform that actually shares data across companies in an open executable standard over time that might very well happen we we we're open to that but it's not where the rates start but also also remember in many cases because if if we if we talk about Conor Coffield or Chevron in one field we compete in all the field we work together and because it you are often joint ventures where we have maybe 10 ownership and chef might have a 10% ownership so in some of these fields they will have access to the shell data or vice versa and where we of course we don't work together of course we only have shell the size with excess I would expect over time that this data gets more and more generally available to be exposed to people who want to develop and melt down for a case for this environment but it's not that's not where it starts okay great question here what is specific to this architecture for oil and gas it looks generic enough that it could apply to any industry which is one of them correct and what will be specific to oil and gas is the or the subsurface space is the wide variety of data source we have which makes it fairly complex because in many industries you're really the majority of the data structure type of data is relatively straightforward here you have many type of untrusted type of data many different data source so you're that's why it puts so much focus on your data ingestion layer your data ingestion layer is so important but as apart from okay we support file types of subsurface there's not the unique about this which is subsurface specific the deliberately did not do that because in principle that we have the same problems in other parts of our industry what is unique if you look at the rest of shell what's unique here is the many data source we have and you see a scale I mean we have more data here than the rest of shell together several times so but can it be used in let's say in the pharmaceutical industry of course it can yeah right which takes us to a question I thought might come up there seems to be a potential natural intersection between this new forum and the open process automation forum has that been discussed at all in your group well I look at opa we're also an active member of opa we're also we're very keen to make sure that opa is successful to make sure we're getting some some new developments in the post control domain so we're very keen on that but we see them as two you you could see that opa will feed into this environment because opa is really your post control domain you're collecting data so is there a link between the two yes there is but I see opa as a feeder into this data platform and it can be data for it can be just time series data can be other types of data so yes in entities always are linked because we use both environments in our industry now when it assume both are successful okay is the drilling real-time in real-time drilling in scope for the forum that's why we talked about real-time data so increasingly especially in the drilling space and that's why also that's I mean especially in drilling space you can do so many things there also with machine learning how the real-time drives your bit down in the ground so yes clearly the real-time piece an important piece of this and the development of well what we stopped but as this what we stopped today is once the well is in production we put into another data platform but that's really what we define so far right and the suggestion here is there may be some real-time api is coming out of the open post automation forum that means we can use them yeah it will be used to use on top of the data platform later then yeah um how much of the subsurface data is proprietary to the individual companies and how are you thinking of addressing the risks of defining a common data architecture for a company's proprietary data and the data might be because of course if we shoot seismic it's our seismic but the formats are not proprietary and so if we if we if we if we ask a company do you go into the Gulf of Mexico and shoot seismic in a certain space and there's shell data and there's many petabytes of data the formats are not proprietary to shell so if Exxon would shoot seismic in another part of the Gulf of Mexico they will use exactly the same formats of course all the data but the same format so what what counts the other formats and the data types not of course the data you need to shell or owned by shell or owned by Exxon owned by Devon owned by somebody else but the the data types are not unique and therefore my focus all the time on the data ingestion layer but that's really what the difference will be made okay um somewhat go a bit a specific question would that would the mining industry benefit from this as well other in yeah I mean what I show over here the data platform every industry is moving into that space yeah and you because if we look at the rest of shell if we go to our retail business do something similar yes but the scale is totally different it's much and much small it's far more structured data the mining where we spoke to one of the mining companies before we start about nine months ago in in Australia of course they have similar data types over there yeah you got similar they yes in principle you can do you could use exactly the same structures out there as I made before it's not unique which is said this is subsurface only yeah what is unique to this so what is what is unique is your are your data types there therefore your data ingestion layer you have your date because in data you do data extraction your you open a file of course that file is maybe unique to the date to the all-in-gas industry but the mining if they have a data structure you would fit their own money so uh from what I know and I'm not the expert in data mining because we shelled it out of mining many years ago um yes it will be uh not to this it will be fairly similar okay question here that's probably more for me than you is can existing members of the open group join this forum or is there an additional charge the answer is this is a new forum so it's a silver membership activity so a separate charge in that sense it is within the gold member entitlement so for those of you who are gold members there's no additional charge that's the answer and platinum of course for those platinum members in the room uh there are many users of subsurface data do you envisage use of uh the open sdw platform in adjacent markets like earthquake tracking geosite yes the same answer yeah because like in the case of shell we already do we already have because people won't know over here but we have earthquake problems in the Netherlands where our one of our last guest foods over there we do exactly that yeah right okay because in that today we're putting sensors out there we're collecting the real-time data and try to predict these earthquakes which will all be successful but of course you're collecting the same data yeah right and that that's it for the questions but there's a general encouragement like we had from Venkat earlier about participating in the in the working group any of you who think your organizations might be interested in this this is just kicking off still in foundation stage and they'd they'd love to have you participate and and help shape this and i look forward to working with you all and the the open group staff and the members of this to give it the value that you're hoping to get from yes from coming here so okay thank you thank you thank you