Seeking to shed new light on Einstein's scientific theories, the Event Horizon Telescope (EHT) project deploys...
hundreds of HGST Inc. helium drives to stream petabytes of image data of Sagittarius A, a black hole at the center of the Milky Way galaxy.
More than 200 of HGST 6 TB Ultrastar HelioSeal hard disk drives (HDDs) are in production use for EHT research. The HGST drives are hermetically sealed with low-density helium to reduce mechanical flutter and power consumption.
The type of drives used by the EHT team is crucial to the research. Several other hard drive brands failed before the group implemented HelioSeal drives. In one installation, 28 of 32 hard drives failed under the extreme conditions. Researchers also considered building a pressured chamber to house its entire computing system.
After data is recorded, the HGST drives are shipped by commercial jetliner to a custom-designed central computing cluster for processing and analysis.
Over time, the HGST helium drives will be used to compile the first image of "the one place in the universe where Einstein's theory of gravity might break down," said Shep Doeleman, an MIT professor who heads the project.
"The vast amount of data we capture completely outstrips normal, run-of-the-mill instruments. This has become a big data issue for us. We got in touch with HGST when we learned the helium drives are hermetically sealed and could withstand the high altitudes," Doeleman added.
EHT is a collaborative effort of 34 observatories and universities. The term, event horizon, refers to a shadowy environment surrounding black holes -- a so-called point of no return, from which light itself does not escape.
Exploring the galaxy gives new meaning to big data storage
EHT scientists use a technique known as very long baseline interferometry to capture the data. Up to 10 radio telescopes situated at various points on the globe are combined to form a virtual radio dish "as large as the Earth itself," Doeleman said.
The network of telescopes reportedly can resolve objects with 2,000 times more granularity than the Hubble Space Telescope. Radio waves are recorded simultaneously from different geographic locations and streamed at a rate of 16 Gbps to the HelioSeal platters.
EHT conducts experiments in scheduled five-day sessions, with each session producing an average of 5 PB to 8 PB of data. Its storage needs to perform reliably at altitudes up to 15,000 feet.
EHT uses the HGST drives in a specially designed two-bay chassis by ConduIT Corp. Each chassis contains eight HelioSeal HDDs configured in a RAID array for 96 TB of raw storage. The project developed software in-house to stripe data across disks for redundancy. Doeleman said his team will add 8 TB HelioSeal drives for greater capacity.
Despite its massive data collection, the EHT project is not yet using traditional backup.
"At the moment, we don't make duplicate recordings of the data," Doeleman said. "If we have a problem at one of the telescopes, we will typically stage another observation at a later time. Our turnkey systems at the sites have a pretty good record for resiliency and robustness. But, as we start to record even more data, we may need to look at adding redundancy."
Examining the environmental impact of helium-filled drives
Toigo: Helium drives don't increase capacity as much as we think
What's in store for the future of helium HDDs?