R2PC: fault-tolerance made easy
Manso, Oscar (1999) R2PC: fault-tolerance made easy. PhD thesis, Dublin City University.
Full text available as:
Fault-tolerance is a concept that is becoming more and more important as computers are increasingly being used in application areas such as process control, air-traffic control and communication systems. However, the construction of fault-tolerant software remains a very difficult task, as it requires extensive knowledge and experience on the part of the designers of the system.
The basics of the Remote Procedure Call (RPC) protocol and its many variants are a fundamental mechanism that provides the adequate level of abstraction for the construction of distributed applications and release the programmers from the burden of dealing with low level networking protocols. However, the standard definition of the protocol does not provide us with semantics that are sufficiently transparent to deal with unexpected hardware and software faults, i.e. the programmer has to deal with possible problems that may occur. To deal with this problem, different reliable variations of the RPC protocol have been defined.
This dissertation introduces a new reliable protocol - R2PC - with the following characteristics.
• Symmetric treatment of client and server processes.
• Use of concurrently processed nested calls in stateful servers.
• The achievement of failure transparency at the application level.
Archive Staff Only: edit this record