August 10, 2022 - Podcast

Episode 295 — The 'holy grail' of data protection

Advances in artificial intelligence and big data analytics rely on data sharing, which can be impeded by privacy concerns. Using a $9 million grant from the National Science Foundation, IU researchers are leading a multi-institution effort to understand how to protect data shared across distributed computing systems such as cloud computing environments. Nearly $3 million of the overall grant will go directly to IU. Computing data has three lifecycle stages: data at rest, data in motion and data in use. Data at rest has reached its destination and is not being used, such as stored data. Data in motion is en route between a source and destination, such as an email on its way to your inbox. Data in use is currently being accessed, read or updated, such as an open Excel spreadsheet. Data at rest and data in motion are typically encrypted for protection in case they're stolen. But data in use is typically unencrypted and therefore more vulnerable to cyber threats. Lead Researcher XiaoFeng Wang, a professor of computer science, engineering and informatics at IU, says data-in-use protection is considered to be a “holy grail” of data protection because even encrypted data has to be decrypted before it can be analyzed, so there’s a risk that the data couple be exposed at that point of time. Wang says the researchers project will lay the technical foundations for practical data-in-use protection across today and tomorrow’s cloud and edge computing systems. He says the effort is critical for maintain U.S. leadership in AI and data science, which relies heavily on data-in-use protection. Led by IU, the project will establish the Center for Distributed Confidential Computing in collaboration with researchers from Purdue University, Penn State, Carnegie Mellon University, The Ohio State University, Spelman College, Duke University and Yale University. The researchers will leverage recent progress in the "trusted execution environment" hardware capability in modern computer chips to run secure computation in a way that can't be compromised by malicious software across distributed computing systems. They will work to provide solutions for data in use, such as training machine-learning models on private data across cloud and edge systems. Wang says the center is the NSF’s largest investment in confidential computing and demonstrates IU’s leadership in cybersecurity research.