Updated: Oct. 18, 2021 at 1:36 p.m.
Researchers in the geography department said they’ve been unable to access their research data for more than a month as officials have been working to transition the Columbian College of Arts and Sciences’ online file storage system to an updated platform.
Officials said they’ve been working to transfer files from the previous, outdated system to the new storage platform since late last year but have found that two volumes of research data that the department uses have “possible data corruption,” blocking access until their transfer. Department members said the technical difficulties have delayed federally funded research, coming at the cost of taxpayer dollars and limited graduate students’ efforts to apply to higher degree programs.
“We understand the importance of these files to our faculty and the impact to their profession as well as the University and are genuinely exploring all of our options to restore as much as possible,” Interim Chief Technology Officer Jared Johnson said in an email.
Johnson said the CCAS file storage system, the CCAS cloud, is “at the end of life” after eight years, and officials have been working since late last year to transition the files to a new platform by December. He said faculty reported that they couldn’t access the data in August, though the data had not been corrupted and was not missing.
“We worked with the vendor to troubleshoot and perform system updates, but there was no improvement in performance,” he said.
Johnson said IT staff worked during mid-September to migrate all faculty and departmental data to a “new storage solution,” with the exception of 80 research volumes that were too big to migrate using the same procedure. Since then, Johnson said IT staff “successfully migrated” 78 of the 80 research volumes in early October to a new storage unit and restored full access to that data.
“We did not include these volumes in the larger migration due to this finding,” he said. “Issues within these two volumes are likely what has been causing the performance issues for the entire system.”
Johnson said the two remaining volumes had “possible data corruption,” and IT staff met with faculty members who rely on the two volumes of data to understand how to prioritize their efforts to restore the data volumes. He said 18 people have access to the two remaining volumes.
“We continue to provide updates and will work with them to identify any missing files due to corruption, or anything GWIT can do to regain access to any missing information,” he said.
IT requests more than doubled during the first week of classes this semester, and faculty reported that slow response times to their requests are the result of “overworked” IT staff. Professors attributed the issues to officials’ move to a shared service model for IT last summer when they consolidated GW’s technology services department to reduce costs and mitigate the financial impact of the pandemic.
The Columbian College Cloud is currently housed inside the University’s Division of IT inside the Foggy Bottom Data Center.
“The primary objective of the Cloud project is to enable Columbian constituents to work in a manner that is most productive, but balance end user flexibility with cost, security and supportability,” according to the University website.
Nikolay Shiklomanov, a professor of geography and international affairs, said he hasn’t been able to access his research data and course syllabi stored on his research drive in the Columbian Cloud since late August.
“On our side, everything has collapsed, starting late August,” he said.
Shiklomanov said the incident has forced faculty and graduate students in the department to call the safety and reliability of the cloud into question, given how the drive has been inaccessible in recent weeks
“The main question is, what is going to happen in the future?” he said. “Because can we trust the system, or can we not trust the system?”
Shiklomanov said his research on the Arctic is funded by a federal grant from the National Science Foundation, and the delays to his work may have come at the cost of taxpayer dollars.
“It’s hard to evaluate the financial impact, but it did give us a little bit of, you know, some scare,” he said.
Michael Mann, an assistant professor of geography, said faculty in the geography department were informed of the possible data corruption, which he says “should be impossible” given that the University stores data in two different locations, Foggy Bottom and Virginia Science and Technology Campus data centers.
Mann said as of Thursday night, IT staff have restored some of their files, but because the files were lost at random, some faculty and students were disproportionately affected by the inaccessible data. He added that this semester’s technical difficulties have cost the geography department “weeks if not months” worth of time.
“I personally have lost confidence in our cloud infrastructure and plan to migrate my data to a more secure platform,” Mann said. “I hope other faculty are at least aware of the risks they face and plan to keep their own backups for their most important work.”
Sonia Clemens, a second-year graduate student in the geography department, said she had to pause her research on permafrost thaw damage in countries near the Arctic Circle for more than a month since the semester began. She said she had to “remake” an entirely new database since she couldn’t access the data she spent the spring semester compiling for her thesis on the geography of Antarctica.
“The work that I spent six months doing, I then had to do in two weeks,” she said.
Clemens said officials sent her an email on Oct. 4 saying they fixed the inaccessible cloud and that she could access her data to continue her thesis research more than a month after the semester began. She said it was the first time officials had communicated with her about her inaccessible data since the start of the semester.
Clemens said the inaccessible data delayed her preparation for a geography conference coming up in October, where she would present part of her research. She said PhD program applicants are expected to share “publications and presentations” from their master’s work.
“That was a little bit scary, feeling that I might not have done adequate work, due to restrictions from technical issues, to continue on to a PhD program,” she said. “That was probably the most frustrating part of like, how could that impact my future just based on a simple technical issue?”
This post has been updated to correct the following:
The Hatchet incorrectly reported that Shiklomanov is an associate professor. He is a full professor. We regret this error.