Hazards of the Cloud: Data-Storage Service’s Crash Sets Back Researchers

Dedoose, a cloud-based application for managing research data, suffered a “devastating” technical failure last week that caused academics across the country to lose large amounts of research work, some of which may be gone for good.

SocioCultural Research Consultants, the company that sells Dedoose, is still scrambling to recover as much of its customers’ work as possible, and has said in a blog post that “the vast majority” of research data on its platform were not affected.

The crash nonetheless has dealt frustrating setbacks to a number of researchers, highlighting the risks of entrusting data to third-party stewards.

Margaret Frye, a postdoctoral researcher at Harvard University, is one of them. Ms. Frye, a sociologist, was using Dedoose to sort and annotate journal entries about AIDS and sexual attraction in southern Africa. But when she logged in on Tuesday evening, she was greeted by a message saying the system was down.

At first, Ms. Frye did not panic. “It wasn’t at all clear that there were data lost,” she said. But when Dedoose finally did come back online, about 60 of the annotated texts were missing.

Ms. Frye and a fellow researcher still had the source material saved on their computers, but more than 100 hours’ worth of “coding”—marking up the journals with metatags and marginalia—had been lost. Ms. Frye is not holding her breath about whether they will be salvaged.

Jayson Richardson, an associate professor of education leadership at the University of Kentucky, had a similar experience.

Mr. Richardson and two of his students have been using Dedoose to “code” job advertisements for elementary- and secondary-school principals, in an effort to see how well Kentucky is preparing its aspiring school administrators to meet the hiring criteria of potential employers.

“In the last two, three weeks,” he said, “we’ve been going at it pretty heavy.”

When Dedoose went down, Mr. Richardson and his colleagues lost about 100 hours of work.

Other testimonials by angry researchers have cropped up on Twitter and Facebook. Many have criticized the company for what they see as a violation of their trust. “Looks like you managed to destroy my wife’s $10,000 research project,” wrote one commenter on the product’s Facebook page.

Eli Lieber, president of SocioCultural Research Consultants, declined to be interviewed about the problems on Sunday, saying he and his colleagues were busy dealing with the problem. He wrote a blog post last week summarizing what had happened.

“This devastating system ‘collision’ of Tuesday night resulted from one of our system-critical Microsoft Azure services failing unexpectedly and, thus, pulling all of Dedoose down,” wrote Mr. Lieber. “The timing was particularly bad because it occurred in the midst of a full database encryption and backup. This backup process, in turn, corrupted our entire storage system.”

For many academics, the transition to storing and managing data using cloud-based services seems inevitable. Universities are increasingly entrusting their data storage to outside companies, and the rise of consumer services like Google Docs has made the cloud seem like a fact of life.

“I’ve been telling my students, Well, this is just the nature of the beast,” said Mr. Richardson, the Kentucky education professor. Every data-storage system can fail, he said, and the benefits of the services available through companies outweigh the risk of storing work on faraway servers.

Still, such incidents can serve as a harsh reminder of that the risk.

“For me, at the time, saving it into the cloud seemed like a way of making my data more secure,” said Ms. Frye, the Harvard researcher.

“But now,” she said, “having this experience, I’m sort of questioning those assumptions.”

Correction (5/12/2014, 3:16 p.m.): This article originally misstated Margaret Frye’s field. She is a sociologist, not a physiologist. The article has been updated to reflect this correction.