A highly decentralized implementation model for theprogrammer-transparent coordination (PTC) scheme for cooperativerecovery

摘要

The authors present an implementation model for theprogrammer-transparent coordination (PTC) scheme that fits well withlocal-area-network-(LAN-) based systems equipped with broadcastingchannels. The model is a significant improvement over the earlierformulated implementation guidelines for developing LAN-basedfault-tolerant distributed computer systems (DCSs). The model uses ahighly decentralized broadcasting-based approach to the execution of PTCfunctions. The result is a significant reduction in the PTC-relatedmessage traffic, and the extent of reduction could be drastic in manyapplication environments. Another major element of the model is athree-layer software structure in which distributed cooperatingapplication processes and PTC-related operating system components areincorporated in modular forms amenable to cost-effective concurrentprocessing