Our group recently began work on a new mobile computing platform for search and rescue and related mapping tasks, in settings where there might be little or no infrastructure communications support. For example, in the immediate wake of the 2005 Tsunami, or on the Gulf Coast after Katrina, affected regions had no communications infrastructure at all, and even weeks later, were limited to intermittent cell telephone and text messaging. Military missions in hostile territory face similar issues. Despite the poor connectivity, one would like to support mobile teams working to search and map the affected region, using lightweight portable devices to capture and record data. These devices will be limited to opportunistic peer-to-peer connectivity, infrequent satellite uplinks, and perhaps some form of very low-bandwidth background communication. Yet we would like to capture data in a secure and trustworthy manner, be able to understand the origins of sensitive data, and ensure that data can't be tampered with or inappropriately disclosed.
In this talk, we'll review the overall problem space and system architecture under development, focusing on the unique mixture of security, attestation and data provenance challenges that arise in this setting. A novel gossip-based communication infrastructure, combined with a more classical group key security architecture, is letting us solve these problems and opening the door to significant improvements in the technology support that will be available to future first-responders.
At Cornell, this work is joint with Dr. Einar Vollset and Dr. Robbert van Renesse. Our work is being performed in the context of DARPA's ACERT program, managed by Dr. Jonathan Smith. Other aspects of the platform are being developed at MIT, Telcordia and Multispectral Solutions.
Gates 4B (opposite 490)