Keeping the cloud secure
EDUCATION, business, government and finance sectors are set to gain a powerful new tool in the race against cybercrime, thanks to a team of Deakin University researchers.
Led by Alfred Deakin Professor Wanlei Zhou, research director of Deakin’s Centre for Cyber Security Research, the team has been awarded a $300,000 Linkage Grant from the Australian Research Council to develop technology that automates privacy processes within cloud-sharing communities across countries.
The researchers are Australia’s foremost experts in the field of “differential privacy,” a concept coined as a result of breakthrough research by Harvard Professor Cynthia Dwork in 2006.
Differential privacy overlaps the areas of statistics and data analytics, and uses techniques such as noise injection to keep the data of individual users completely private.
“This is cutting-edge technology,” Prof Zhou said.
“Differential privacy had its first large-scale practical application last year when Apple used it as a tool for phone input patterns to guarantee privacy. It has huge potential.”
Over the three-year project, the Deakin researchers will also call on machine learning and data mining to develop intelligent systems that can preserve individual privacy and take into account the privacy laws of individual countries — reducing costs, speeding up communication times and improving security.
The project is expected to result in a commercialised data sharing system for the cloud environment.
It will initially benefit educational organisations, but will lay the foundations for data sharing in other cloud communities such as governments, banks, and other industries in Australia in the coming years.
The team will also include Deakin’s Dr Tianqing Zhu and Dr Gang Li, in partnership with one of Australia’s largest international edu- cation companies, Australian Education Management Group (AEMG).
AEMG has established cooperation between educational institutes and research centres in China, the US and Europe. In China, it co-ordinates more than 30 joint programs involving 12 Australian universities, including Deakin joint programs with Southwest University and Inner Mongolia Normal University.
Prof Zhou said the system would be adopted by AEMG’s cloud campus to exchange data in a restricted privacy manner between partner institutions.
The technology will be commercialised as a “middleware” that can be plugged into existing cloud environments to maintain required privacy — even when the cloud crosses various jurisdictions with different privacy policies.
“Currently, there is no real automation available for privacy preservation,” Prof Zhou said.
“Different countries have their own privacy policies, and sensitive documents have to be manually checked before they leave a country.
“We will develop a method to achieve differential privacy over time, so adversaries can’t identify the information through their own data mining or machine learning tools.
“There are several traditional methods for ensuring privacy currently, but they have limitations. The use of encryption requiring passwords limits access to data to a very small number of people, while the use of anonymity reduces the value of data.
“If the machines of adversaries have enough background information, they can identify individuals over time anyway. We will be able to prevent these issues with a much more sophisticated technique.”