Full width home advertisement

Welcome Home


Post Page Advertisement [Top]

NASA Turns to the Cloud for Help With Next-Generation Earth Missions



New, cutting-edge Earth science satellites that will be launched within a couple of years will provide us with unprecedented views of our home planet. We'll be able to track small-scale ocean features such as coastal currents, which transport nutrients essential to marine food webs, as well as the amount of fresh water flowing through lakes and rivers, which will help us better understand the ocean. We'll also be able to detect changes in the Earth's surface that are less than half an inch in size (a centimeter). Although these satellites will generate a deluge of data, engineers and scientists will be prompted to develop cloud-based systems capable of processing, storing, and analyzing all of the digital information generated by these satellites.


'About five or six years ago, it became clear that future Earth missions would generate a massive amount of data for which current data storage systems were inadequate,' said Suresh Vannan, manager of NASA's Physical Oceanography Distributed Active Archive Center in Southern California. "The systems we were using would quickly become insufficient," he added.

 

In NASA's Earth Science Data Systems program, the center is one of several that are tasked with the responsibility of processing, archiving, documenting, and disseminating data collected by the agency's Earth-observing satellites and field experiments. Over the course of several years, the program has been addressing the information-volume challenge by migrating its data and data handling systems from on-premises servers to the cloud – software and computing services that run on the internet rather than on a user's computer.

 

However, even though the amount of data returned by the spacecraft is not as large as that returned by many future satellites, the Sentinel-6 Michael Freilich satellite, which is a component of the US-European Sentinel-6/Jason-CS (Continuity of Service) mission, was the first NASA satellite to make use of this cloud system.

 

One hundred terabytes of data will be generated every day by the SWOT and NISAR missions, which are both scheduled for launch in 2018. A terabyte is approximately 1,000 gigabytes of digital storage space, which is equivalent to approximately 250 feature-length films' worth of digital storage. Southwest Oceanography and Topography (SWOT) will generate approximately 20 terabytes of science data per day, while the NISAR mission (National Aeronautics and Space Administration-Indian Space Research Organization Synthetic Aperture Radar) will generate approximately 80 terabytes of science data per day. The Physical Oceanography Distributed Active Archive Center will archive SWOT data, whereas the Alaska Satellite Facility Distributed Active Archive Center will archive NISAR data, according to the Department of Commerce. NASA's current Earth science data archive contains approximately 40 petabytes (1 petabyte equals 1,000 terabytes), but the archive is expected to grow to more than 245 petabytes by 2025 – just a few years after the launch of SWOT and NISAR – a significant increase from its current size of 40 petabytes.

 

It is anticipated that both NISAR and SWOT will rely on radar-based instruments to collect data. With a planned launch in 2023, NISAR will monitor the Earth's surface, collecting data on environmental characteristics such as land shifts caused by earthquakes and volcanic eruptions, changes to the Earth's ice sheets and glaciers, and changes in agricultural activity and vegetation cover.

 

In addition to measuring the height of the planet's surface water, both ocean and freshwater, SWOT will assist researchers in compiling the world's first survey of fresh water and small-scale ocean currents, which will be launched in 2022 and will be the first of its kind. Collaboration with NASA and France's space agency, the Centre National d'Etudes Spatiales, is taking place to develop the SWOT analysis.

 

According to Kevin Murphy, NASA's Science Mission Directorate's Chief Science Data Officer, "This is a new era for Earth observation missions, and the massive amount of data they will generate will necessitate a new era for data management." In addition to facilitating efficient access to a shared cloud infrastructure across the agency, NASA is also training the scientific community on how to access, analyze, and utilize that data.

 

The Earth science satellites currently transmit data to ground stations, where engineers convert the raw data of ones and zeroes into usable and comprehendible measurements. The raw data is processed, which increases the file size. However, for older missions that return relatively small amounts of data, this is not a significant issue. After that, the measurements are transmitted to a data archive, which stores the information on computers called servers. The majority of the time, when a researcher wants to use a dataset, they go to a website, download the data, and then work with it on their computer.

 

For the vast majority of scientists, however, this will be impossible because of missions such as SWOT and NISAR. Using SWOT to download a day's worth of data onto a computer would necessitate the purchase of twenty laptop computers, each with a storage capacity of one terabyte. Using an average home internet connection, it would take approximately a year for a researcher to download four days' worth of data from NISAR. Working with cloud-based data eliminates the need for scientists to purchase large hard drives or to wait months for a large number of large files to download to their computer system, saving them money. According to Lee-Lueng Fu, project scientist for SWOT at the Jet Propulsion Laboratory, "By processing and storing large amounts of data in the cloud, we will be able to study big-data problems in a cost-effective and efficient manner."

 

As a result, organizations will no longer be required to pay for the storage of massive amounts of data or the physical space required to house all of those hard drives, which will relieve them of infrastructure constraints. According to Hook Hua, a JPL science data systems architect for both missions, "we simply do not have the additional physical server space at JPL with sufficient capacity and flexibility to support both NISAR and SWOT."

 

NASA engineers have already taken advantage of this feature of cloud computing to develop a proof-of-concept product based on data from the Sentinel-1 satellite. In addition to NISAR, the satellite is part of an ESA (European Space Agency) mission that studies changes to the Earth's surface, though it does so using a different type of radar instrument than NISAR will use. Engineers used Sentinel-1 data in the cloud to create a colorized map of the Earth's surface, depicting everything from more vegetated areas to deserts. According to Paul Rosen, project scientist for NISAR at JPL, "it took a week of continuous cloud computing with the equivalent of thousands of machines to complete the task." "Had you attempted this outside of the cloud, you would have had to purchase all of those thousands of machines," says the author.

 

In Earth science, cloud computing will not replace all of the methods by which researchers interact with scientific datasets, but according to Alex Gardner, who works on NASA's NISAR science team and studies glaciers and sea level rise, the technology is gaining traction. He anticipates that in the near future, the majority of his analyses will be performed off-site rather than on his laptop or personal server. According to him, "I fully anticipate that in five to ten years, I will have very little space on my computer's hard drive and will be exploring the cloud's new firehose of data."

No comments:

Post a Comment

Bottom Ad [Post Page]