Unification of replication and transaction processing.

摘要

This dissertation describes a novel software architecture that provides robust fault tolerance for networked applications within and between enterprises. The architecture resolves many of the problems currently associated with the use of transactions, and roll-backward recovery, for networked enterprise applications. In particular, the dissertation focuses on the transparent replication of applications that are built on top of Commercial-Off-The-Shelf (COTS) distributed transaction processing middleware and the Common Object Request Broker Architecture (CORBA) standard for distributed object computing.; With the unified architecture that we have developed, transactions are used locally in each enterprise to protect the application data, while the application logic, its processing and its interactions with computers in other enterprises are protected by replication and roll-forward recovery. Consequently, an application never retracts any data transmitted from one enterprise to another, which greatly simplifies the application programming and makes networked enterprise applications more robust.; The core of the unified architecture is the Pluggable Fault Tolerance (PluggableFT) infrastructure. It is built using the pluggable protocols framework that is available for most CORBA Object Request Brokers (ORBs). The pluggable protocols framework allows for seamless integration of the replication mechanisms into the ORB and, thus, facilitates a fault tolerance solution that requires minimal modification of the ORB and of the application program. The integration makes it possible to develop sophisticated mechanisms to replicate and recover transaction processing applications, which are typically very complex.; In-depth performance evaluation of PluggableFT and of the unified architecture has been conducted. The measurements and the analysis have led to a number of optimizations of the fault tolerance infrastructure and have led to the idea of how to structure large-scale fault-tolerant applications for the best performance.