GGrantIndex
← Search

CT-ISG: Diaspora: A Secure, Reliable Federated Execution Platform and Object Store

$357,125FY2006CSENSF

Cornell University, Ithaca NY

Investigators

Abstract

NSF-0627649 TITLE CT-ISG: Diaspora: A Secure, Reliable, Federated Execution Platform and Information Store PI Andrew Myers It is difficult to coordinate computation and to manage information across organizational boundaries and wide-area networks. To address this problem, this project develops a distributed, federated platform for secure, distributed computation over consistent, durable objects. The platform is a universal system for managing structured information and securely computing on it, allowing allow very large numbers of objects to be linked together on behalf of many mutually distrusting users, while giving users control over where objects are stored. This enables organizations to cooperate in a fine-grained, low-cost way, with benefits for many activities including education, medicine, business-to-business transactions, corporate and military information systems, and large-scale cooperative information resources. The project goes beyond previous distributed systems platforms in several ways. It provides integrated support for both distributed storage and distributed computation, whereas previous platforms have focused on one or the other. It supports cooperative sharing of information and computation even when the cooperating entities do not fully trust each other. And it aims to automatically enforce a larger class of desirable system properties: not just security or availability narrowly defined, but also properties such as consistency, confidentiality, integrity, efficiency, and durability. Replication, encryption, and distributed commit protocols make it possible to automatically ensure the consistency, integrity, and availability of persistent information. A high-level language with security and consistency annotations enables programmers to easily write code for this system, using a compiler that transforms code and data to produce a secure, reliable, and efficient system.

View original record on NSF Award Search →