FT-RC4: A robust security mechanism for data stream systems

Stream database systems are designed to support the fast online processing that characterizes many new emerging applications such as sensor-based environments, online business processing, and network monitoring. Data stream processing is a highly demanding environment where streams are usually infinite, bursty, and run at high arrival rates. Due to limited buffer storage or real-time constraints, data items may be dropped out of the system and lost forever. In many applications, sensitive stream data needs to be secured against malicious attacks. Various security mechanisms have been well studied in the literature. However, these mechanisms are not tuned to work in the lossy streaming environment. Stream security mechanisms are required to provide security services and to be fault-tolerant as well.

In this paper, we identify the security requirements for data stream systems, focusing on Nile, a prototype query processing engine for data streams developed at Purdue University. We first propose a security architecture for data stream systems, then focus on a particular service: data integrity and confidentiality. We present a new mechanism, FT-RC4, that provides data integrity and confidentiality. We demonstrate its practicality by implementing it inside our prototype data stream system and evaluating its performance.