As computers become more sophisticated, researchers are generating more data than ever before. On the whole this is a terrific development, but it has created an interesting challenge: where do you store all the data?
UCLA’s Institute for Digital Research and Education offers storage with its high-performance computing, but it’s fairly expensive and not easily available outside of the cluster environment. Consequently there is a growing campus demand for inexpensive, reliable archival research data storage. To solve this latest challenge, IDRE technologists are developing a project that would provide a low-cost, state-of-the-art cloud storage system to UCLA researchers.
“Lots of people have data they want to store but there isn’t a performance requirement attached to it,” said IDRE chief technologist Scott Friedman, who conceived the plan. “They just need to put it somewhere. Sometimes they go to a Big Box store and buy a storage drive. But that means they’re putting valuable data on unreliable storage. It could be intellectual property that belongs to the university, or it could have taken a lot of time to generate. You shouldn’t put it on a drive, which you might drop, break, or lose.”
Recent changes in government grant programs requiring potential funding recipients to document how they plan to store and share data have also contributed to the need for a centralized storage network, Friedman added.
Cost was a top priority in planning the system, because the service needs to be inexpensive enough that researchers will want and can afford to use it. In order to achieve this outcome, Friedman worked backwards, starting with the price, and then contacting vendors to see if they were willing to meet it.
“I went to vendors and told them what I needed and how much I could pay for it,” Friedman explained. “I told them if they were willing to sell it to me for that amount I would have people using it and I would buy more. To a company they said, `OK, no problem.’”
The final numbers are still being crunched, but Friedman is confident the service will cost less than a storage drive from a Big Box. And it will provide researchers with far greater security.
“We’ll be providing safe, reliable storage for research data at a better price than people could get on their own, so they really will have no excuse not to use it,” he said. “It actually will be significantly cheaper per unit of storage. And we’ll be offering it through an interface that is convenient for them.”
The system will be simple and autonomous. As conceived, researchers essentially will go to a web portal, purchase an amount of storage, and provide a recharge number. The system will do the rest, including automatically billing a researcher’s account. Researchers will be able to interact with their stored data through their personal computers. Any replication will be taken care of behind the scenes – or, if you prefer, the clouds.
While evaluating demand for a campus-wide cloud storage system after coming up with the idea in April 2012, Friedman was enthusiastically encouraged by IDRE Research Technology Group Director Bill Labate, who immediately endorsed the concept. UCLA senior leadership has been enthusiastic, as well.
“The response was universally positive,” Friedman said. “If only I had a dime for every time I told people about this and they said, `Can I buy it right now?’”
Pilot testing is set to begin in November, and the target date for campus-wide implementation is spring of 2013.