Amazon Becomes a Large Cog in the Industry’s Digital-Transformation Machine
DSDE recently spoke with Bill Vass, vice president of engineering for Amazon Web Services, about his observations on the oil and gas industry’s digital efforts and Amazon’s aggressive growth in the business.
Digitalization is taking hold in the oil and gas industry, and the world’s largest tech companies are at the center of it. Among them is Amazon, whose work in the industry took off earlier this decade when operators and service firms began migrating to Amazon Web Services’ (AWS) cloud computing platform. DSDE recently spoke with Bill Vass, vice president of engineering for AWS, about his observations on the industry’s digital efforts and Amazon’s aggressive growth in the business.
Vass, a petroleum geologist by training, is AWS’ executive sponsor for the oil and gas space. During his early years, he paid for college by working on offshore seismic vessels and began his career at Western Geophysical in Houston writing software for oil and gas applications. After serving as president and chief executive officer of Liquid Robotics, he joined AWS in 2014 and has overseen teams developing storage, automation, and monitoring products for a variety of large enterprises, including those in the oil and gas industry.
What opportunities have emerged for AWS as the oil and gas industry digitalizes en masse?
“We’ve always felt that the cloud is an extremely good fit for oil and gas. It has huge amount of petaflops [available for] seismic processing and reservoir simulations. It has zettabytes of storage available for seismic data and other large amounts of data. It’s highly secure, highly reliable, geographically distributed, and—best for oil and gas—it’s elastic, and that’s something much of the industry doesn’t have today. If they’re on an on-premises data center, and the price of oil drops, they can’t shut their data centers down and sell them off. And then, when oil goes up again, they can’t buy new data centers fast enough to meet demand. But, with the cloud, you can literally turn it off and on—like a spigot.
“Amazon also has a long history of working with intelligence agencies and the US Department of Defense, which operate in very rough environments all over the world. So we’re used to having durable devices like our Snowball Edge (https://aws.amazon.com/snowball-edge/) transportable compute and storage device that will operate in high and low temperatures and extend the cloud beyond the network. This is ideal for edge processing and machine-learning applications on ships, platforms, and in field locations.”
What are the biggest challenges facing the upstream space when it comes to digitalization?
“I think part of it is a skills gap. This is new technology to a lot of folks. Some people feel that it’s intimidating. It doesn’t need to be. It really isn’t that different from running virtual machines, and your data centers, they are just elastic and available in a different way.
“Another challenge is making sure that the folks using the cloud are sophisticated enough to ask the right questions of cloud vendors. When we say the word ‘Region,’ for example, we talk about three buildings that are 10 to 60 km away from each other where your data is striped across those separate buildings, which are referred to as ‘Availability Zones.’ So you can lose a building and not lose any of your data or application availability. Other cloud providers say the word ‘Region,’ and it’s only one building, so when a tornado takes out the building, you’re out of business.
“Security is another area that’s important for people to understand. Our cloud has been certified at the highest security level there is in the industry by the intelligence agencies for top-secret/[sensitive] compartmented information [TS/SCI]. We got that certification in 2014. We know how to operate in those areas, and we build that security into all of our systems.
“People should also understand the ROI [return on investment] of comparing a capital expense on premises with a flexible, recurring revenue; things like Spot Instances, which are sort of on-demand, low-cost compute; Auto Scaling that allows you to scale up when you have high demand and scale down when you’re not running your large loads, so you’re not paying for computers when you’re not using them; and using the proper storage tiers to save money on colder data.”
Amazon is one of a few tech firms offering cloud services to the oil and gas industry. How is AWS different?
“Like I said, we’re the only ones that have been certified for TS/SCI Level 6 data since 2014, so we have a lot of experience operating high-security environments. We’ve had a ruggedized remote processing device [Snowball Edge] since 2014 that can operate in oil fields and deserts to collect, store, process locally, and ship seismic data—plus operate on wells and platforms.
“We have a really robust IoT [Internet of things] environment, a really robust robotics and automation environment. Those are some of the things that differentiate us in many ways. We also are a vendor that offers lots of choice. So other than Oracle itself, we’re the only vendor that you can run an Oracle database on. You can run Microsoft SQL Server on us. You can run Windows on us. You can run all different flavors of Linux on us. We’ve got ARM servers, which are very low cost, based on the ARM chip. We’ve got FPGAs, GPUs, and Intel chips. So we have a lot of variety there.
“We have more varieties of databases. We have more experience than anyone else in machine learning right now and more machine learning running on us than any other vendor out there. We pioneered the serverless space, which is a whole new type of computing. We’re also about 14 times larger than all the other cloud providers combined. So we have the breadth and scope to handle any technology processing load in the oil and gas industry.”
Can you provide any case studies in which AWS has improved upstream work?
“We’ve done work with Shell more recently on the Open Subsurface Data Universe where we created a data lake. Shell took many, many years of their data and used machine learning to cleanse it and load it into this data lake running on our S3 platform. They built a complex identity management system for limiting access to that based on a person’s role with the company. Then, they created a geospatial interface for it, a graphic interface where you can draw a trapezoid anywhere in the world and Shell’s assets pop up. And they’ve also done things like use machine learning to read in old data such as well logs, then clean and digitize them. Once the well logs and seismic data are all in digital form, machine leaning can be used to correlate rock formations across well logs and seismic data.
“Woodside has done both upstream and downstream work with us. They’re preprocessing seismic onboard ships with Snowball Edge devices and shipping the data in and optimizing high-performance computing on AWS for their seismic processing, doing their seismic visualization. You see folks like Hess and Petronas doing the same kinds of things. We were first to support Halliburton platforms.
“Subsurface data lakes, seismic interpretation and processing, reservoir simulation and management, and using drones and autonomous systems for surveys are all very popular.
“We’re working with a bunch of different companies to put video and LiDAR instruments in their operating areas and then apply health and safety machine-learning algorithms to that video to keep their folks safe. We’re also using it for simple things like knowing whether the service or water truck arrived or not—regardless of whether the company was actually billed for it—and for warning people on the North Slope that polar bears are approaching their rig.”
As some countries move to keep their data within their borders, how is AWS handling the complicated issue of data residency?
“We’ve got Regions all over the world—in most of the places that really matter. We’ve got a Region in the Middle East that’s for data residency across the United Arab Emirates and Saudi Arabia. We have Regions in China, and data residency is obviously important there. We have Regions in Canada, the United States, India, Africa, and all over Europe.
“Generally speaking, we meet most of our data-residency requirements just by operating in those places, along with the fact that we have very strong cryptography, very strong security, and we meet their audit requirements.
“There are a few places where we don’t yet have Regions. And, in those places, some countries have decided that it’s okay to go ahead and put it nearby as long as you’re controlling the encryption keys locally in-country. For some of those countries, we put in Snowball Edge clusters, which allows us to store the data and operate it locally. And we have a new product called Outposts, which will allow them to keep the data in-country for places where we don’t yet have a Region. But, at our current Region growth rate, I really think there’ll be very, very few places where we won’t be able to keep the data in-country if that’s the requirement.
“There will be a lag for some places. For example, Nigeria and Russia right now are two places where we don’t have Regions. The Nigerian data may be able to be stored in our South African Region—we’re still discussing it with the government. If not, we’ll put an Outpost there with local data storage. And Russia will probably have Outposts with local data storage as well.
“We also have over 200 points of presence all over the world that allow people to connect in quickly with low latency into a private fiber network.”
Is there anything else you’d like to say to oil and gas firms that are embarking on their digital journeys?
“I think the most important thing I can stress is training as they’re going down this digitalization path. Demystify the cloud, get your internal IT folks trained, get some of your business executives trained. Because then you can make informed decisions and know what’s fud [fear, uncertainty, and doubt] from what’s real—you understand how the ROIs work, you understand how the cloud works, you understand the breadth of everything available in the digital space. “
Oh, and is AWS hiring?
“We are very rapidly expanding. The demand for oil and gas far exceeds our ability to deliver right now. There are many, many more projects that we’re working on than we have professionals.
“We’re always looking for people with a good technical acumen, people in oil and gas, people who are customer-focused because we’re a very customer-focused company given our retail background. So they know what a mud log is, they know what a well log is, they know the difference between a log measuring resistance vs. an X-ray log and things like that. We want, for example, folks with experience with rig automation, and then also folks experienced on the retail and health and safety sides.
“So we’re just hiring like crazy right now in the oil and gas space because it’s a very big market for us. We’re very excited about the market. Oil companies are very innovative in their operations and how they do things. They’re early adopters of technology in a lot of domains. So it’s exciting to work with them in the high-performance computing and cloud space as they innovate.”