WSU developing efficient, cost-effective method for outsourcing large-scale data crunching and computer storage
DETROIT - As today's businesses struggle to meet their needs for large-scale information processing and storage, one Wayne State researcher is working to make outsourcing these services a more efficient, cost-effective option.
Song Jiang, assistant professor of electrical and computer engineering in the College of Engineering and resident of Troy, Mich., recently received a $400,000 Faculty Early Career Development (CAREER) Program award from the National Science Foundation to develop a more efficient method for quantifying Quality of Service (QoS), a guaranteed level of performance in data flow. The technology will be developed for large IT companies including Microsoft, Google, and Amazon.com, who offer their cloud computing and storage services with their extensive IT infrastructures to organizations looking for more computing power and data storage space.
Although a popular idea, outsource computing and virtual storage services have not developed an efficient method of quantifying hard disk input/output QoS - a parameter needed to predict the price for service. "Hard disks are mechanical, and the information on them is stored much like a record," Jiang said. "Depending on what information is needed, the time spent locating the information can vary. This inconsistency makes it difficult for companies that outsource their data storage and processing services to specify service quality and set a rate for services."
To solve the problem, Jiang is proposing the implementation of consolidated storage service based on a reference storage device, which can be the user's own system. The user is then guaranteed QoS that is as good as their own system, regardless of variations and changes in data access patterns.
Jiang's research is of high interest to IT giants such as Yahoo!, who use very large scale consolidated storage as an infrastructure for mission critical applications such as Yahoo search engines. For his research, Jiang is receiving technical support from Yahoo! as well as the Performance and Architecture Laboratory at Los Alamos National Laboratory, a United States Department of Energy national laboratory, and Seagate Research, a manufacturer of hard drives and storage solutions.
"As today's computing relies more and more on efficient and reliable access to large amounts of data, this research has the potential to significantly improve the productivity and efficiency of many IT services," Jiang said. "It will be important both for mid-size companies outsourcing their computing and virtual storage as well as very large organizations looking to build or expand their own computing facilities."
Jiang's research will also serve as a source for virtual data and storage system training for his student assistants and curriculum for WSU computer and electrical engineering students. "We're excited not only to pave the way for one of the most relevant technologies of the future, but also address the concern for the lack of strong data and storage system professionals in the IT industry."