CC* Data: ImPACT - Infrastructure for Privacy-Assured compuTations

点击翻译按钮获取中文摘要

This project addresses one of the fundamental problems in distributed computing: the ability to compute on sensitive or private data in a secure manner. This is critically important for multi-institution and cross-disciplinary sharing and analysis of data, particularly in health-related fields and the social sciences. This project builds upon and enhances several cyberinfrastructure elements which have been developed through other NSF awards and are already part of the existing research base. By developing methods for securely sharing private data in different collaborative settings, ImPACT (Infrastructure for Privacy-Assured CompuTations) lowers the barriers to using privacy-restricted data in scientific collaborations. It reduces the time to discovery and opens new avenues for research.ImPACT enables a variety of research scenarios that require integration of data across subdomains where relevant data sets are held by different stakeholders. The goal of ImPACT is to enable cooperative processing across the stakeholder-owned datasets, while respecting the privacy policies of the individual owners, and to provide a model for collaboration that could be readily used by other institutions. The project brings together a team that includes social science researchers, cyber-infrastructure experts, and distributed systems and security experts from the University of North Carolina, Duke University, and Indiana University. A non-profit organization formed by the city of Durham, NC provides some of the data sets that are used in the project. To enable end-to-end data flows, ImPACT builds upon prior investments by the NSF Campus Cyberinfrastructure program at the Duke and UNC campuses, as well as the Global Environment for Network Innovations (GENI) program to support straightforward transfers of data within the respective virtual infrastructures. By supporting three different trust models between collaborators, the project develops diverse methodologies with best practices in networking, data management, security, and privacy preservation to accommodate a variety of collaborative scenarios. Data access, sharing, processing, and publishing are integrated into privacy-aware systems that allow scientists to use their own tools and that build upon enabling cyberinfrastructure technologies like Dataverse, CyVerse, and the Open Resource Control Architecture (ORCA) control software.