Replica Voting Based Data Collection: Adaptive Protocols and Application Studies

Data collection in a hostile environment requires dealing with malicious failures in the sensing devices and underlying transport network: such as data corruptions and message timeliness violations. Functional replication is employed to deal with failures, with voting among the replica devices to deliver a correct data to the end-user. The paper describes the protocol-level design issues for voting with low message overhead and delivery latency under various failure scenarios. The well-known 2-phase voting protocol is employed as a building-block, with different protocol variants dynamically selected based on the network conditions and device-level fault severity. A case study of replicated web services is also presented to illustrate the usefulness of our dynamic adaptation mechanism