In recent years, data security has been more information-centric rather than device-centric. A trusted information system ought to provide secure and fault-tolerant delivery of sensor data to the end-user applications even though the physical sensor devices and components may suffer failures or get attacked by malicious adversaries. In an example of multi-vehicle cooperative driving system for autonomous cars, a malicious sensor can possibly mis-report about the presence of road obstacles with an intent to to cause collisions. A solution
approach is the functional replication at device and/or software levels, which allows enhanced fault-tolerance and security in the sensor data collection infrastructure.

My talk shall focus on how data and software can be replicated in a secure sensor-based information system, and the engineering and operational constraints therein. Specifically, replica voting as a means to ensure high integrity data collection from sensor devices will be discussed. Some case studies from our CUNY research on data replication software systems and Information-centric networks will be presented: namely, collaborative autonomous driving in intelligent highways and multi-UAV coordination for surveillance applications.