Danielle Applestone: When was the last time you took a sea voyage? Maybe never, but odds are good that many, perhaps most of the things you own have. Ships carry 90% of the world’s goods, more than 70% as containerized cargo. Figures show that carrying freight by shipping container can be 17 times more fuel efficient than transporting those same goods by air and 10 times more than by road. A large cargo ship might have an approximate dead weight, the total weight in cargo, stores, and bunkers that it can hold, of 220,000 tons or about the weight of the Sears Tower in Chicago. As you can imagine, it takes a fair amount of fuel to push this ship from port to port. On a given day, say today, the world’s shipping industry might consume 5 million barrels of fuel and account for two and a half percent of the planet’s CO2 greenhouse gas emissions. The good news is technology is on the case. A San Francisco-based tech firm called Spire is leveraging on-demand high-performance computing resources in the cloud to help transform worldwide shipping. By using information to help cut fuel consumption, they’re helping reduce greenhouse gas emissions and cut costs in a sector whose margins are often paper thin. What’s more, their ability to gather weather data, with unprecedented speed and detail, provides ships with real-time information that can help them avoid disaster. In the next few minutes you’ll learn that the secret to Spire’s data isn’t out at sea, or in one of their six offices on terrafirma…
Speaker 2: Two, one, we have liftoff.
Danielle Applestone: But in space, orbiting hundreds of miles above the nearest vessel. I’m Danielle Applestone, engineer and entrepreneur. And though I haven’t captained a super tanker yet I did row crew back in college. This is Technology Powers X. In this episode, Technology Powers, Smooth Sailing. Aboard the MV Wakashio, July 25th, 2020. In calm seas, the Japanese oil tanker is sailing at 11 knots near the coast of Mauritius, a little too near as it turns out.
Speaker 3: Authorities are investigating a recent oil spill…
Speaker 4: A ship has run aground on a coral reef off the Indian ocean.
Speaker 9: The surrounding coral reef causing an ecological emergency.
Danielle Applestone: The Wakashio ran aground on a reef where less than two weeks later it would begin to split in two, leaking a thousand tons of oil into the Indian Ocean. Why? Constantine Komodromos suggests it was a failure of technology. He’s one of the co-founders of VesselBot, a company that provides digital solutions for the international maritime industry.
Constantine Komodromos: The crew could not get good connectivity to use internet and telephone on board and they diverted from their planned route. However, could not see the real feed of data where the vessel was, they never knew that this vessel was going out of their normal route. This caused the vessel to go and run aground in sea because they went very close to shore and that created a lot of issues both in regards to costs to remove the vessel, and then it broke down into two and there were pollution issues.
Danielle Applestone: VesselBot uses technology that helps ships avoid disasters such as that of the Wakashio. It uses complex data to make shipping safer, more economical and better for the environment.
Constantine Komodromos: VesselBot offers a voyage optimization and GHG emissions reduction system where users, shipowners and charters can use in order to identify the route, the speed over ground and the RPM that engine should have in order to move from location A to location B.
Danielle Applestone: Sailing from A to B is becoming more complicated and more crowded. Since the 1990s worldwide vessel traffic has increased four fold, helping explain that 2.5% share of greenhouse gas emissions and shrinking profit margins. That’s where companies like VesselBot come in, using data to ease safety and environmental concerns and to dial back fuel costs, and an increasing amount of that data is coming from space. Which brings us back to Spire whose network or constellation of a hundred-ish small satellites is helping prove that the immediate future of space is getting smaller.
Danielle Applestone: Spire was founded in 2012 and fueled at first by a Kickstarter campaign. It’s small band of space-minded entrepreneurs were inspired by the Cube Satellite pioneer Bob Twiggs, whose small satellite revolution began as a class project for his students at Morehead State University in Kentucky. It was Twiggs who democratized space by designing and deploying tiny homemade cube satellites inspired by the cubic four-inch acrylic boxes used to package Beanie Babies. Given that it costs about $10,000 to put one pound of payload into orbit, the small satellite movement has made space accessible and affordable for a new generation of tech entrepreneurs. The more the Spire team molds small satellites, the more advantages they discovered, for instance, using software to constantly upgrade the technology from earth.
Tom Henderson: The benefit of reprogrammable radios in orbit is first of all, you don’t have to get everything perfect before you launch the satellite.
Danielle Applestone: Tom Henderson is a Senior Software Engineer at Spire Global.
Tom Henderson: You can upgrade your software once you’re on orbit based on lessons you learned in orbit. And that really lowers our costs. In the traditional satellite industry you’ve got to get everything perfect before you put that thing up there and this allows us to increase performance even if there’s launch delays. So even if we’re not launching satellites our constellation can be getting better and better.
Danielle Applestone: Old school satellites were often the size of a bus, costing billions of dollars and requiring more than a decade to design and build. Once in orbit it can only survey a portion of the earth at a time. A network of 100 Spire mini satellites can be built and deployed faster and at a fraction of the cost, covering more of the planet.
Tom Henderson: Since we can get a lot more satellites up there that are small, we can put them in a lot of different orbits and we can get much larger coverage of the earth’s surface for the same cost as a large satellite that can only see one spot at a time.
Danielle Applestone: In space, size, weight, and the cost of repair can make for big problems. When the Hubble Space Telescope initially failed to work properly, repair missions were launched at a cost of $700 million. Spire’s small satellites on the other hand are inexpensive and replaceable. Put in another way, failure is an option.
Tom Henderson: At the moment there is no cost-effective way to repair small satellites like ours in space. It just costs too much to even send a robot up there to try to do something with them, it’s cheaper to launch another satellite.
Danielle Applestone: The team at Spire discovered several advantages to deploying their small in low earth orbit.
Tom Henderson: One of the reasons is that they can be small because they don’t need a lot of power to transmit and receive data. But the other nice fact of low-Earth orbit is that there is still tiny wisps of atmosphere up there. And what that means is that the lifetime of a satellite in low-Earth orbit is measured in a handful of years, so even if your satellite breaks and is completely useless it’s eventually going to re-enter the atmosphere and burn up so we don’t have this endless problem of space junk accumulating like you do in a geosynchronous orbit, that’s much higher.
Danielle Applestone: Some 500 miles below, the shipping industry is already reaping the benefit of Spire’s data.
John Lusk: What our data ultimately allows our customers in the maritime world to do is really threefold.
Danielle Applestone: John Lusk is Vice President of Global Data Services at Spire.
John Lusk: We can help track vessels. So it’s just from a pure safety perspective and understanding perspective we can help shipping companies know where their vessels actually are, we can help in terms of fuel efficiency, so really understanding the speed at which vessels are moving, estimated time of arrivals. We can use all kinds of different datasets to understand how we’re getting the most efficiency out of the fuel that we’re spending, which drives really more efficient on spending on fuel. We can also use that data to track weather.
Danielle Applestone: That’s where a responsive constellation of satellites can be lifesaving, conveying fast, detailed weather data to a ship that may be heading into trouble.
John Lusk: If we infuse our weather data with our maritime AIS data we can start to develop a routing prediction. For instance, we’re moving 10 knots to the West, but oh wow there’s a massive storm coming in and we predicted the winds from the storm are going to be 30 knots, we should maybe go a little bit further south and then up the coast. We can actually provide that level of detail and really help shipping companies understand how to avoid storms and which will not only save costs but also potentially save lives.
Danielle Applestone: Traditional real-time weather data from sources such as Great Britain’s Shipping Forecast comes from balloons sent up to test the atmosphere. Spire satellite constellation records radio occultation, revealing detailed atmospheric data. Where weather balloons are typically launched by humans in populated areas, no such barriers apply to a space-based satellite constellation, effectively democratizing weather data. John Lusk.
John Lusk: Whereas somebody in the middle of Bangladesh, if you’re looking for a weather forecast in the middle of Bangladesh you’re probably not going to get a very accurate forecast because no one’s really focused on that particular part of the world. They might have a forecast for a major city in Bangladesh but where you are in the middle of nowhere, you’re probably not going to get a great forecast, but with our radio occultation in our weather data infused into our weather model you can now get an accurate forecast no matter where you are in the world.
Danielle Applestone: And that includes the middle of the ocean. A proud moment in Spire’s young life came in 2020. As daily routines turned upside down, air travel was dramatically reduced and that created a data gathering problem.
John Lusk: Unbeknownst to me and most of us at Spire was that a big part of weather forecasts, in fact, one of the biggest data inputs to having a weather forecast is data that you get from an aircraft. Every airplane whether it’s commercial or private that flies up into the air has a sensor on it that collects all kinds of atmospheric data. That data is sent to NOAA and ECMWF and a lot of the other world-leading weather service providers that are out there. And so it’s a massive data input to helping deliver an accurate weather forecast.
John Lusk: When COVID hit in March all of a sudden all the planes are grounded, uh oh, there’s no more aircraft data and that had a massive impact on the accuracy of weather forecast around the world. In fact, I think it was something like 13 or 14% decrease in the accuracy of weather forecast, so NOAA and ECMWF which is the European leader in weather forecasting were kind of freaking out. They were trying to figure out, “What are we going to do?” And they knew about radio occultation, in fact, that had been one of their smaller data sets that they’d been using to fuel their weather forecast.
John Lusk: But Peter Platzer, our CEO came up with this brilliant idea of like, “You know what? This is our opportunity to really help them out. We can help these weather providers make sure that they’re delivering an accurate weather forecast with our own data because if you’re not delivering accurate weather forecasts that can be catastrophic, not just individuals, people, countries, societies, but also for companies.” And so we offered our data for free to ECMWF, they ingested our data over the course of a month and they found out that not only was our data more accurate than the aircraft data it was displacing, but it actually reduced the risk of having an on-earth dataset.
Danielle Applestone: Meanwhile, back at sea, space-generated data is helping cut fuel consumption for shipping companies in several ways. Detailed weather data allows ships to plot the most fuel-efficient course. Data can factor in shipping congestion at a port, calculating the most efficient time and route for departure and arrival. Analytics can factor in geographic barriers, icebergs, and the type of ship to calculate the safest, most fuel-efficient route and speeds. To Constantine Komodromos of VesselBot, such technology renders a disaster like the Wakashio completely avoidable.
Constantine Komodromos: We’ve used historic data from Spire and these billions of waypoints and data points, we collected those data points from Spire and we used machine learning and artificial intelligence in order to be able to build a network of different waypoints that could be connected to each other in order to route one vessel from one location to the other.
Danielle Applestone: For the Spire team, the only place to amass weather data was poetically enough, the cloud, and on-demand HPC cloud. Tom Henderson.
Tom Henderson: With the exception of laptops and monitors our entire IT environment is in the cloud. In fact, when I joined the company, Peter Platzer the CEO said, “Hey, Henderson, here’s your laptop. Go build 24/7 operational global numerical weather prediction and I’m not spending a dime on capital investment for you. So go figure out how to do it all in the cloud.” And I can’t resist challenges like that so of course I had to do it, and with a great team I’ve been able to hire here we’ve accomplished that. And for us it’s been a big deal because when I joined the company four and a half years ago we didn’t know of anybody else who’ve actually accomplished this so I have to say it’s been a lot of fun.
Danielle Applestone: An on-demand, turnkey, HPC environment proved to be a perfect fit for Spire and no small part because of its speed. Garima Kochhar is a distinguished engineer on the HPC engineering team at Dell Technologies.
Garima Kochhar: If you think of an organization like Spire which has an interest in weather forecasting, high-performance computing methods allows them to ingest a lot more data and to solve more complicated problems in a shorter period of time. Specifically for weather simulation or weather forecasting, if the complex equations that I use to describe physical phenomenon about tides, and atmosphere, and wind, and temperatures and so on, if the amount of time it takes me to compute what all these things mean for tomorrow’s weather, if that computation takes three days to do then by the time I get the result the result is no use to anybody.
Danielle Applestone: Tom Henderson describes the technology that powers Spire data.
Tom Henderson: We use Dell systems from both of our main HPC as-a-service vendors right now. R Systems, where we’re running our 24 seven operations. We’re actually using M1000 chassis because they work in the configuration that R Systems has, it worked very well for us. For R&D at Ohio Supercomputing Center, we’re using a lot of Dell hardware for their Owens supercomputer where they have a PowerEdge C6320s. And then they just have been putting in a new machine called Pitzer and it’s got like 658 PowerEdge 6420s with a mix of Skylake and Cascade Lake processors.
Danielle Applestone: Turnkey HPC resources, delivered as-a-service, enables Spire to avoid the cost and appreciation of buying an in-house fixed size resource, and it’s far better suited to the fits and starts nature of their data usage.
Tom Henderson: The way numerical weather prediction works is that for us at least four times a day we need to make a forecast and that forecast needs to complete as fast as possible because of weather forecast is another thing where the value decreases rapidly with time. What that means is that we have a very bursty usage pattern. Four times a day we’re going to use everything we can get our hands on, and in between those times we’re not going to use anything so 24/7 ownership of a fixed size resource it’s just too costly for us.
Danielle Applestone: For Garima Kochhar, this convergence of satellite data and high-powered computing couldn’t come soon enough.
Garima Kochhar: When you think about weather today, you can see how much we’re able to plan for, but how much more there is still to do where we are in reactive mode. And techniques like high-performance computing, artificial intelligence, and data analytics and companies like Spire with their innovation and their leadership show us how much more that can be done, whether it’s in weather modeling and forecasting and predictions that improve shipping routes, that improve our ability to forecast natural disasters, that allow us to have insight to address climate change.
Danielle Applestone: John Lusk notes that the success of Spire is as much a story of the right timing and the right technology as it is about its scrappy startup attitude.
John Lusk: The fact that we’re smaller and we’re able to maintain that startup mentality is a benefit. It allows us to move faster, we don’t have the corporate processes that a lot of the bigger space companies do, for instance. We are able to become a bit more nimbler, we can make decisions without going through nine levels of approval. We had an autonomous nature within the company where the folks that we hire are able to just motivate themselves and get things done and they’re not afraid to jump in and just make something happen, they’re not afraid to take risks and they’re not afraid to fail fast. It’s not about failing per se, it’s really about what are you learning when you do fail and are you able to take those learnings and share them with the rest of the company and the other individuals so that they don’t make those same mistakes?
Speaker 2: Three, two, one…
Danielle Applestone: Among those learnings for Spire is that the more time they spend in space, the more possibilities emerge. Their aptitude for successfully deploying small satellites has attracted companies, asking them to share their secrets. For them, Spire has launched an orbital services branch, call it insights for hire. For other clients, Spire provides help interpreting the data they provide. Meanwhile, its technology is enabling new weather insights related to soil moisture and wind speed on the ocean surface and may help revolutionize the enforcement of illegal fishing. “All I ask,” wrote poet John Masefield, “Is a tall ship and a star to steer her by.” 120 years later the great ships have found a new star to guide them. Information, data, drawn from the skies bringing smooth sailing, cleaner air, and a lot less drama on the high seas.
Danielle Applestone: This is Technology Powers X, an original podcast from Dell Technologies. For more information on Dell Technologies HPC solutions, go to DellTechnologies.com/hpc. For more information on Dell EMC PowerEdge servers go to DellTechnologies.com/servers. And to learn more about this episode, our speakers, and to read the transcript, visit DellTechnologies.com/technologypowersx. I’m Danielle Applestone, thanks for listening.