Reliable Processing in Distributed Databases

本项目研究分布式数据库中可靠处理的算法和模型。重点是多站点故障造成的问题，这些故障加上丢失的消息、通信软件中的拥塞（可能导致网络分区）以及部分复制的数据库系统的一致恢复。该研究将整合在事务处理背景下正在开发的几个概念，特别是提交/终止协议和并发控制理论。

This project investigates algorithms and models for reliable processing in distributed databases. The focus is on problems created by multiple site failures coupled with lost messages, congestion in communication software (that may cause network partition), and consistent recovery of partially replicated database systems. The research will integrate several concepts that are being developed in the context of transaction processing and in particular the commit/termination protocols, and concurrency control theory.