Contributed by Guy Holmes, written for the recently published Volume 4, Issue 1 of the PPDM Foundations Journal for 1Q2017.
The technologies that are changing the IT landscape around us often take time to get traction in the oil sector. Sometimes it appears we, as an industry, are so heavily invested in yesterday, that tomorrow is too much to think about. But sometimes revolution does not start within the disquiet, rather it starts next door and spreads into the open arms of disquiet as though it had been waiting for it all the time.
Some ground breaking disruptive technologies that are appearing on the landscape within an array of other industries will at some point seek to pass on their benefits in the oil sector. How long it takes to adopt them is up to us.
This article speaks at a high level to the impacts of a small but important group of these technologies including Hadoop, Block Chain, Internet of Things (IoT) and IBM’s Watson and how these technologies will likely one day be an important part of our industry – open arms or not.
Several years ago, I attended a course put on by the Australian Institute of Company Directors (AICD). The course focuses on the areas that all company directors need to be aware of in order to be a well-rounded, responsible and professional company director. It covers areas like financial literacy, legal responsibilities, strategy, etc. A few months ago, I attended the course again as a refresher so that I could get myself up-to-date on the latest developments in company director issues and responsibilities.
One of the major differences between the first course and the latest update was that it included a significant section on business risk in the area of “Disruptive Technology”. It is a review of the risks that companies face when it comes to emerging technologies that can make sweeping changes to an entire market with little to no notice. It reviews ways to identify these technologies, participate in these developments, and how to protect a company’s interests from these disruptions.
Like the car for the horse, or the Uber for the taxi, disruption development has become an industry in its own right, with developers hunting for ways to make tools and services that will cut through an industry and grab market share in massive chunks. What follows is a review of new technologies that stand to improve (and potentially disrupt) current oil industry technologies and practices.
HADOOP
What is it?
Hadoop is a programming framework that provides tools for the processing and storage of extremely large datasets in a distributed computing environment. The framework is open source and a product of The Apache Software Foundation (www.apache.org).
Why is it important?
Hadoop has made significant impacts in cost savings and efficiencies where (in particular) large datasets are involved. The method used by Hadoop to store very large files is a distributed file system that can map data wherever it sits. Hadoop can also use the local resources of where the data sits to process the data so that the processing can be done in the same location. In addition, Hadoop can also handle data that is unstructured meaning that data does not necessarily need to be in rigid tables within a database. Hadoop effectively lets you see everything you have from one spot, process it from many, and removes the need to “over-manage” data through databases, tables, etc.
What are some example adaptations?
Hadoop should not be looked at as a replacement for your current technology, but rather a toolset to augment and improve it. There are many uses for Hadoop in the oil sector. I am sure that some of these have already been put to good use by the more innovative companies in our industry. But for those that have not, it is worth giving this technology a good hard look.
An Example:
An explorer wants to look at some seismic data from the Gulf of Mexico to answer a question about a particular geology. The dataset required is 20Tb in size and is currently on tapes that need to be read to disk to be processed. In a typical workflow, the 20Tb will be read and processed in sequentially. In the end, the explorer creates about 200Mb of new data from this work and spends 2 weeks in doing so.
Hadoop, in the first instance, makes keeping volumes of data (like this 20Tb) in a highly available state and it reduces the traditional costs of storage because it removes a lot of the overheads of managing that storage. Hadoop then allows the dataset to be processed in situ. In many cases in the oil industry, data has to be copied to where the processing will take place. Hadoop, however, allows the processing to take place wherever the data is, i.e. instead of bringing the data to the processing, the processing can come to the data.
With the right tools in place, the example above would be reduced from weeks (most of which waiting for data to be made available and processed) to just several hours. Hadoop can work with structured data like massive databases or unstructured data like miscellaneous file types, including; log files from web sites, spreadsheets, social media interactions, incoming sensor data, and streaming field and production data.
BLOCKCHAIN
What is it?
A blockchain is essentially a distributed database that continuously maintains a list of ordered records called blocks. Each block contains a timestamp and a link to a previous block as well as other metadata about that block. A distributed database is a database which is not attached to a single common processor and, in fact, may be stored on multiple computers, located in dispersed locations across the globe, spanning an array of interconnected computers. Blockchains are inherently protected from modification after a block has been written as once recorded, the data in a block cannot be retroactively altered. Blockchain is the backbone of the management of the digital currency Bitcoin.
Why is it important?
Aside from blockchains use in the expanding market of digital currency, it has numerous applications where the need to track (without error or possible altering of information) is required. Just as blockchain works within the currency market, so can it be used to track physical items like art, music, a barrel of oil or a software license. Although PPDM does not use a blockchain model, PPDM should consider the implications of blockchain and how this could impact the next generation database model.
What is an example adaptation?
In the seismic acquisition market, speculative acquisition companies are constantly acquiring new data and licensing this data to oil companies for use in their exploration programs. In many cases, the licensing of the data they acquire has contractual requirements that prevent the purchaser from transferring the data to another party, or if transfer is permitted, it usually has uplift or transfer fees attached. Currently, acquisition companies rely on the oil companies, press releases or market intelligence to notify them of data transfers. Vast sums of money are not recovered by acquisition companies every year because they are not aware of data transfers that have occurred.
A blockchain could be used to track every seismic trace in a survey acquired by an acquisition company, the licensing of each trace in that survey, and any transfer of that trace from one party to another. Embedding blockchain technology into a “smart contract” for the sale of the trace data, for example, would allow an oil company to transfer the trace data to a new oil company and the blockchain, in turn, would automatically track that change, automatically debit the bank account of the oil company and automatically credit that of the original licensor.
INTERNET OF THINGS (IoT)
What is it?
The Internet of Things is broadly considered the networking of physical devices that enable these objects to collect and exchange data. To better imagine what the term internet of things might encompass, picture a series of temperature sensors located in the ocean placed around a gas production platform. Information is collected by these devices in real time and is continuously sent back to the platform to monitor ocean temperature around the platform. Now connect that platforms sensor array to all of the other platform arrays in the Gulf of Mexico and then connect the Gulf of Mexico’s array to those deployed globally by a weather forecasting agency. Then connect all of these to the International Space Station which is monitoring average ocean temperatures in the northern hemisphere in real time to analyse global warming trends. Imagine then that the international space station could then send commands to all of these sensors to cool or heat in response to real time changes it sees from the sensors.
Why is it important?
IoT is all around us now in our everyday lives. From its use in smart buildings to monitor air conditioning, to getting traffic alerts on a smartphone when driving, IoT is providing life changing intelligence on scales large and small.
What is an example adaptation?
IoT is already in use in the oil sector but has a lot more to offer as the technology (and our imagination) improves. Imagine an oil company that has both upstream and downstream investments like a producing oil field, a refining operation and 50 retail gas stations. As petrol is sold at the retail gas stations, the plant is notified in real time of consumption trends for unleaded, leaded, high octane, etc. products by sensors on the petrol bowser. The plant can increase or decrease production of the various products to meet the demands of its clients based on this input. In return, the plant can communicate with the production field on its requirements for raw material in an effort to keep production at an optimum level to service the whole supply chain. The efficient use of the IoT can eliminate the need for the expensive storage of refined output until it is consumed, it can prevent lost revenue from gas stations selling out of a product, and can significantly improve the efficiency of an entire supply chain.
IBM WATSON
What is it?
Watson is a question answering computer system that was invented by IBM as part of its DeepQA project. The Watson system is capable of answering questions posed in natural language by accessing structured and unstructured content that has been provided to it in advance to build its knowledge base. IBM’s goal is to develop computers that interact in natural human terms across a range of applications such as the fields of medicine and law. To achieve this, IBM needed to program the computer system to understand the questions that humans ask, in the many ways that humans can ask them, and provide answers that humans can easily understand.
Many of us probably think that this technology is already all around us in systems like Siri or voice activated navigation systems. But the reality is that most of these systems only respond to a limited array of questions that have to be phrased in a certain way on a very specific subject matter. Watson on the other hand has been designed to understand complex questions and phraseology posed to it in a wide range of grammatical contexts. It does this by parsing questions into sentence fragments to identify phrases that are statistically related and then searches its knowledge base to locate possible answers to the questions. It then ranks all of the possible answers and uses algorithms to determine which one is the best possible fit as the answer to the question. It may not seem like a significant advancement, but in many ways it is ground breaking.
As an example, there is a well-known joke by Groucho Marx that goes something like “One morning I shot an elephant in my pajamas”. The sentence could easily be interpreted in several different ways. For example, who was wearing the pajamas – the person or the elephant? When he shot the elephant, did he use a gun or a camera? Fortunately Marx finishes the joke with, “How he got into my pajamas I’ll never know” so we then understand the context.
Watson has been programmed to break down these phrases and create context to allow it to interpret the sentence in a way a human might.
Why is it important?
Aside from the fact that the world has been trying to replace humans with machines for a very long time, the impact of Watson can and will be profound in many other ways. Whether it is a doctor or nurse asking Watson about a cancer treatment plan, or an investor trying to understand what influences the price of oil, Watson is already a ground breaker in the area of artificial intelligence.
What are some example adaptations?
Like most companies, often the knowledge base for a specific area in the business rests inside the head of the last person who worked on a project. Watson is changing that dynamic by ingesting any and all data that can be obtained from that specific area of the business, organising it, remembering past questions, the feedback you provide it, and not only presenting its answers in a meaningful way, but also refining the answers as new information becomes available.
Watson can be used in exploration, for example, to monitor the progress of a drilling project and help prevent expensive downtime on the rig. By feeding Watson live inputs from the drilling progress like current lithology, current depth, current temperature, type of drill bit, and the rotational speed of the bit, Watson could compare all other wells drilled 
in that basin, or through that type of geology, with that type of bit and at that speed to determine if maximum drilling efficiency is being maintained. It could alternatively be used to determine if there is the danger of a looming mechanical failure by monitoring trends and historical failures for tell-tale signs of impending breakages. Watson can do this by consistently reviewing the inputs it is provided, looking at historical drilling projects and reports from the past, analysing the drilling contractors experience and the drill rigs specific toolsets, and provide meaningful intelligence about the project that a human simply would not be able to provide.
Woodside Petroleum in Australia implemented Watson in 2015 to help on a range of fronts, including; infrastructure construction, production monitoring, and predicting future trends from historical data and after a year of use, Woodside is increasing its use of the technology across a broad range of business units.
Combing These Technologies
Aside from using each of these tools independently, it won’t be long before most or all are used in conjunction with one another. It is not hard to imagine an array of sensors in an IoT creating unique records that are tracked in a blockchain, processed by an implementation of Hadoop and then streamed into a data repository that forms the knowledge base for Watson to ingest. With a combination like this, you could get real time highly qualified and useful data, processed at high speed, emanating from a remote device that has full provenance and ownership tracked.
It will take considerable vision to come up with solutions that combine these tools in an end to end system. But whoever does will be onto something quite special.