Adaptive QoS Mechanisms for Dependable Data Delivery in Replication-based Voting Systems

摘要

Majority voting among replicated data collection processes enhances the trust-worthiness of data flowing from a hostile external environment. It allows a correct data fusion and dissemination by the end-users, in the presence of content corruptions and/or timing failures that may possibly occur during data collection. In this paper, we provide a QoS-oriented approach to manage the data flow through various system elements. The underlying protocol-level parameters that influence data delivery performance are the data sizes, network bandwidth, and voter asynchrony. A replica voting protocol takes into account the interplay between these parameters as the faulty behavior of malicious processes unfolds in various forms during data collection. Our QoS-oriented approach casts the well-known fault-tolerance techniques, namely, 2-phase voting, with control mechanisms that adapt the data delivery to meet the end-to-end constraints - such as latency, miss rate, and resource cost. The QoS-oriented approach allows us to infuse the currently prevalent network-level mechanisms for resource management into the voting protocol operations in a seamless manner, while meeting both the fault-tolerance and end-to-end delivery performance.